BTC_Predictor.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "BTC Predictor.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU",
    "gpuClass": "standard"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "# Get Data"
      ],
      "metadata": {
        "id": "78Y7lQmv9QuF"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Importing time series data with pandas"
      ],
      "metadata": {
        "id": "RBNRMCisATuG"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "df = pd.read_csv(\"/content/drive/MyDrive/BTC Predictor/BTC_USD_2013-10-01_2021-05-18.csv\", \n",
        "                 parse_dates = [\"Date\"],\n",
        "                 index_col = [\"Date\"])\n",
        "df.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 237
        },
        "id": "pZT0pfKS_638",
        "outputId": "67160310-c423-4ef8-c322-ed6af0665d3f"
      },
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "           Currency  Closing Price (USD)  24h Open (USD)  24h High (USD)  \\\n",
              "Date                                                                       \n",
              "2013-10-01      BTC            123.65499       124.30466       124.75166   \n",
              "2013-10-02      BTC            125.45500       123.65499       125.75850   \n",
              "2013-10-03      BTC            108.58483       125.45500       125.66566   \n",
              "2013-10-04      BTC            118.67466       108.58483       118.67500   \n",
              "2013-10-05      BTC            121.33866       118.67466       121.93633   \n",
              "\n",
              "            24h Low (USD)  \n",
              "Date                       \n",
              "2013-10-01      122.56349  \n",
              "2013-10-02      123.63383  \n",
              "2013-10-03       83.32833  \n",
              "2013-10-04      107.05816  \n",
              "2013-10-05      118.00566  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-c5180e3c-3e80-41b0-a515-0254661cdd39\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Currency</th>\n",
              "      <th>Closing Price (USD)</th>\n",
              "      <th>24h Open (USD)</th>\n",
              "      <th>24h High (USD)</th>\n",
              "      <th>24h Low (USD)</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Date</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2013-10-01</th>\n",
              "      <td>BTC</td>\n",
              "      <td>123.65499</td>\n",
              "      <td>124.30466</td>\n",
              "      <td>124.75166</td>\n",
              "      <td>122.56349</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2013-10-02</th>\n",
              "      <td>BTC</td>\n",
              "      <td>125.45500</td>\n",
              "      <td>123.65499</td>\n",
              "      <td>125.75850</td>\n",
              "      <td>123.63383</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2013-10-03</th>\n",
              "      <td>BTC</td>\n",
              "      <td>108.58483</td>\n",
              "      <td>125.45500</td>\n",
              "      <td>125.66566</td>\n",
              "      <td>83.32833</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2013-10-04</th>\n",
              "      <td>BTC</td>\n",
              "      <td>118.67466</td>\n",
              "      <td>108.58483</td>\n",
              "      <td>118.67500</td>\n",
              "      <td>107.05816</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2013-10-05</th>\n",
              "      <td>BTC</td>\n",
              "      <td>121.33866</td>\n",
              "      <td>118.67466</td>\n",
              "      <td>121.93633</td>\n",
              "      <td>118.00566</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-c5180e3c-3e80-41b0-a515-0254661cdd39')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-c5180e3c-3e80-41b0-a515-0254661cdd39 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-c5180e3c-3e80-41b0-a515-0254661cdd39');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df.info()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OjdTBzpPAMyb",
        "outputId": "a3e9eb9f-eb24-4fac-d2c2-84addb1ccdb7"
      },
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "DatetimeIndex: 2787 entries, 2013-10-01 to 2021-05-18\n",
            "Data columns (total 5 columns):\n",
            " #   Column               Non-Null Count  Dtype  \n",
            "---  ------               --------------  -----  \n",
            " 0   Currency             2787 non-null   object \n",
            " 1   Closing Price (USD)  2787 non-null   float64\n",
            " 2   24h Open (USD)       2787 non-null   float64\n",
            " 3   24h High (USD)       2787 non-null   float64\n",
            " 4   24h Low (USD)        2787 non-null   float64\n",
            "dtypes: float64(4), object(1)\n",
            "memory usage: 130.6+ KB\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#How many samples?\n",
        "len(df)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "WLsttbd4FBvd",
        "outputId": "706be2e2-027b-44a5-a2a4-60bbc6fa23fa"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2787"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#Only want closing price for each day\n",
        "\n",
        "bitcoin_prices = pd.DataFrame(df[\"Closing Price (USD)\"]).rename(columns={\"Closing Price (USD)\": \"Price\"})\n",
        "bitcoin_prices.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 237
        },
        "id": "fp7t1cPhFGqQ",
        "outputId": "04d73af3-f633-49fc-bb70-674eabb97ad4"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                Price\n",
              "Date                 \n",
              "2013-10-01  123.65499\n",
              "2013-10-02  125.45500\n",
              "2013-10-03  108.58483\n",
              "2013-10-04  118.67466\n",
              "2013-10-05  121.33866"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-4b7da66f-fa53-410c-8e64-9f25179134d7\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Price</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Date</th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2013-10-01</th>\n",
              "      <td>123.65499</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2013-10-02</th>\n",
              "      <td>125.45500</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2013-10-03</th>\n",
              "      <td>108.58483</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2013-10-04</th>\n",
              "      <td>118.67466</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2013-10-05</th>\n",
              "      <td>121.33866</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-4b7da66f-fa53-410c-8e64-9f25179134d7')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-4b7da66f-fa53-410c-8e64-9f25179134d7 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-4b7da66f-fa53-410c-8e64-9f25179134d7');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "bitcoin_prices.plot(figsize=(10,7))\n",
        "plt.ylabel(\"BTC Price\")\n",
        "plt.title(\"Bitcoin Price from 1 Oct 2013 to 18 May 2021\", fontsize=16)\n",
        "plt.legend(fontsize=14)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 478
        },
        "id": "RzWc0L4vHxZp",
        "outputId": "6bd28b30-9c26-45d8-9cab-176442e18e0d"
      },
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7fc787f8d510>"
            ]
          },
          "metadata": {},
          "execution_count": 10
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x504 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Format Data Part 1"
      ],
      "metadata": {
        "id": "K-eHYl9CEqtH"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Creating train and test splits for Time Series"
      ],
      "metadata": {
        "id": "5LdPpGCvNZWt"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Turn DataFrame Index and column into NumPy arrays\n",
        "\n",
        "import numpy as np\n",
        "timesteps = bitcoin_prices.index.to_numpy()\n",
        "prices = bitcoin_prices[\"Price\"].to_numpy()\n",
        "\n",
        "timesteps[:10], prices[:10]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-flyZS-nN-3F",
        "outputId": "b650bfec-9cea-4b55-98b2-368be0eccace"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array(['2013-10-01T00:00:00.000000000', '2013-10-02T00:00:00.000000000',\n",
              "        '2013-10-03T00:00:00.000000000', '2013-10-04T00:00:00.000000000',\n",
              "        '2013-10-05T00:00:00.000000000', '2013-10-06T00:00:00.000000000',\n",
              "        '2013-10-07T00:00:00.000000000', '2013-10-08T00:00:00.000000000',\n",
              "        '2013-10-09T00:00:00.000000000', '2013-10-10T00:00:00.000000000'],\n",
              "       dtype='datetime64[ns]'),\n",
              " array([123.65499, 125.455  , 108.58483, 118.67466, 121.33866, 120.65533,\n",
              "        121.795  , 123.033  , 124.049  , 125.96116]))"
            ]
          },
          "metadata": {},
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Wrong way to make train/test sets for time series\n",
        "from sklearn.model_selection import train_test_split \n",
        "\n",
        "X_train, X_test, y_train, y_test = train_test_split(timesteps, # dates\n",
        "                                                    prices, # prices\n",
        "                                                    test_size=0.2,\n",
        "                                                    random_state=42)\n",
        "X_train.shape, X_test.shape, y_train.shape, y_test.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "IOjaDLvFILMM",
        "outputId": "9cfb5b64-a2cf-43c9-e309-f9a6185b2e4e"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((2229,), (558,), (2229,), (558,))"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Plotting wrong train and test splits\n",
        "plt.figure(figsize=(10, 7))\n",
        "plt.scatter(X_train, y_train, s=5, label=\"Train data\")\n",
        "plt.scatter(X_test, y_test, s=5, label=\"Test data\")\n",
        "plt.xlabel(\"Date\")\n",
        "plt.ylabel(\"BTC Price\")\n",
        "plt.legend(fontsize=14)\n",
        "plt.show();"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 442
        },
        "id": "1ssCjEzCOrxl",
        "outputId": "835bff24-c938-4dc3-d635-04da6dbed929"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x504 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "This kind of random split is okay for datasets without a time component (such as images for classification problems) but for time series, the time factor has to be taken into account.\n",
        "\n",
        "We need to split our historical Bitcoin data to have a dataset that reflects the past (train set) and a dataset that reflects the future (test set)."
      ],
      "metadata": {
        "id": "hxxznUt5O2A3"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#Create train and test splits the right way for time series data\n",
        "split_size = int(0.8 * len(prices)) # 80% train, 20 % test\n",
        "            \n",
        "#Create train data splits (everything before the split)\n",
        "X_train, y_train = timesteps[:split_size], prices[:split_size]\n",
        "\n",
        "#Create test data splits (everything after the split)\n",
        "X_test, y_test = timesteps[split_size:], prices[split_size:]\n",
        "\n",
        "len(X_train), len(X_test), len(y_train), len(y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EjsdfsjcOsL9",
        "outputId": "922342e9-546c-4086-ad17-97f0c2f422b3"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(2229, 558, 2229, 558)"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        " #Plot correctly made splits\n",
        " plt.figure(figsize=(10,7))\n",
        " plt.scatter(X_train, y_train, s=5, label='Train data')\n",
        " plt.scatter(X_test, y_test, s=5, label='Test data')\n",
        " plt.xlabel(\"Date\")\n",
        " plt.ylabel(\"BTC Price\")\n",
        " plt.legend(fontsize=14)\n",
        " plt.show();"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 442
        },
        "id": "ZjuYnF1kPyq2",
        "outputId": "ba8ef6b1-c6d9-4e4c-de0c-823c3c35ec65"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x504 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAnAAAAGpCAYAAAANygvZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeXxU9b3/8ddnEgLZWBNAdn9XUNBCWYRg1fZKXXC5eqtW7QKoVX+1i7d2ubVat97bxZ/VantrixWXLtqWq7eKdaHU22olYMCVpYpVCAjZQMgCSWbO9/fHOZPMJJNFIJnt/Xw88phzvufMme9AbT58l8/HnHOIiIiISPoIJbsDIiIiIvLBKIATERERSTMK4ERERETSjAI4ERERkTSjAE5EREQkzeQmuwP9raSkxE2aNCnZ3RARERHp0bp162qdc6Ud27MugJs0aRIVFRXJ7oaIiIhIj8xsa6J2TaGKiIiIpBkFcCIiIiJpRgGciIiISJpRACciIiKSZhTAiYiIiKSZrNuF2h3P89i+fTuNjY3J7oocRgMGDGDkyJEMHjw42V0RERE5LBTAxaitrcXMOProowmFNDiZCZxz7N+/nx07dgAoiBMRkYygKCXG+++/z6hRoxS8ZRAzo6CggLFjx1JdXZ3s7oiIiBwWilRiRCIRBgwYkOxuSB/Iz8+ntbU12d0QERE5LBTAdWBmye6C9AH9vYqISCZRACciIiKSZhTAiYiIiKQZBXDSyZIlSzj77LMP+3OXL1+uqUwREZHDQAFcGjOzbn+WLFlyUM+96667+NWvfnV4O3uQJk2axO23357sboiIiKQU5YFLYzt37mw7XrFiBVdccUVcW35+ftz9ra2tvdplO2TIkMPXSRERkXQVCUPtm1B6DKRYirHU6o18IKNHj277GTp0aFzbgQMHGDp0KA8//DCnnHIK+fn5/PznP6euro5LLrmEcePGkZ+fz7HHHsv9998f99yOU6gf+9jHuPrqq/nWt75FSUkJI0eO5Gtf+xqe53Xbv4ceeoiJEydSUFDA2WefTVVVVdz1t99+m3PPPZfRo0dTWFjIrFmzWLFiRdznbt26la9//etto4pAr76DiIjIIYmE4bYj4Z758IOJ/nkKUQDXBzzPUVPfjHMu2V3huuuu4+qrr2bjxo2cd955HDhwoC1Q2rBhA9dccw1XXXUVq1at6vY5v/71r8nNzeXFF1/kJz/5CT/60Y/47W9/2+X9a9asYcmSJVx55ZW88sornHPOOdx4441x9zQ0NLBw4UJWrlzJq6++yvnnn88nPvEJNm/eDMCjjz7KuHHjuPHGG9m5c2fb6OLBfgcREZFeq30Tmvf5x837/PMUoinUw8zzHJfcW866rXuYPXEYD19RRiiUvIX7X/rSl7jgggvi2r7+9a+3HV955ZX8+c9/5uGHH2bBggVdPmfatGnceuutAEyZMoV7772XVatWcckllyS8/6677mLBggVcf/31be956aWXuO+++9rumTFjBjNmzGg7v/7663niiSdYvnw5N9xwA8OHDycnJ4fi4mJGjx7ddt/YsWMP6juIiIj0WskUsBxwEf+1ZEqyexRHI3CHWV1jC+u27iHsOdZt3UNdY0tS+zNnzpy480gkwn/+538yffp0RowYQVFREY8++ijbtm3r9jnTp0+POx8zZky3pak2bdrE/Pnz49o6njc2NvKNb3yDadOmMWzYMIqKiqioqOixLwf7HURERHpt/+7uz5NMI3CHWUlRHrMnDmsbgSspyktqfwoLC+POb7/9dn74wx9y11138aEPfYiioiK+9a1v9VgntOPmBzPrcQ1cT772ta/x9NNPc/vttzN58mQKCgpYtGgRLS3dB70H+x1ERER6rbAUJpRB5RoYP88/TyEK4A4zM+PhK8qoa2yhpCgv5fKevfDCC5xzzjl89rOfBcA5x5tvvtm2CeJwmTp1KuXl5XFtHc9feOEFFi1axPnnnw/4a9vefvttpkxpH6bOy8sjEokk5TuIiEgWM4PFK6Cp1g/eUuz3uaZQ+0AoZJQWD0y54A38tWirVq3ihRdeYPPmzXzxi1/knXfeOeyf8+Uvf5k//elPfO973+Ott97i3nvv5bHHHuvUl8cee4z169fz+uuv85nPfIYDBw7E3TNp0iSef/55duzYQW1tbb9+BxERyXKhEBSUQGMNpMDGxFgK4LLMDTfcwNy5c1m4cCEnn3wyhYWFfPrTnz7sn1NWVsZ9993HPffcw/Tp03n00Ue5+eab4+654447GDlyJCeddBILFy6krKyMk046Ke6eW2+9lcrKSv7pn/6J0tLSfv0OIiKS5TwPHjwb7pgKD5zln6cIS4VUF/1pzpw5rqKiIuG1TZs2MXXq1H7ukfQX/f2KiMgH0lDtB29eGEK5cO0mKBrZr10ws3XOuTkd2zUCJyIiIpJIYam/gSGUm3IbGbSJQURERCSRFN7IoABOREREpCuhUL9Pm/aGplBFRERE0owCOBEREclunudvWEijjZ0K4ERERCR79ZQqJEWDOwVwIiIikr2aav1yWV7Yf22qbb+Wwnng+jSAM7OhZrbczDab2SYzm29mw81spZm9FbwOC+41M7vbzLaY2WtmNivmOYuD+98ys8Ux7bPN7PXgPXdbKpY+EBERkdQVTRViOTBmpl95Iaq74C7J+noE7i7gaefcMcAMYBPwTWCVc24ysCo4B1gITA5+rgTuATCz4cBNwDxgLnBTNOgL7rki5n1n9PH3ERERkUxiBoseh7Gz4L2X/RG36EhbCueB67MAzsyGACcD9wE451qcc+8D5wIPBrc9CJwXHJ8LPOR85cBQMzsCOB1Y6Zzb7ZzbA6wEzgiuDXbOlTu/nMRDMc+SJLr99tuZNGlSsrshIiLSO/t3+8Fbx5E2LwILb4N/2wBLnkypPHB9OQJ3JFAD3G9mL5vZL8ysEBjlnNsZ3LMLGBUcjwUqY96/PWjrrn17gvZOzOxKM6sws4qamppD/Fqpw8y6/VmyZMlBP/vmm2/muOOOO3yd7YGZsXz58n77PBERkTaJRtoiYbjtSPjZR+C/jveDuRTSl4l8c4FZwJecc2vM7C7ap0sBcM45M+vzbR3OuaXAUvBrofb15/WXnTt3th2vWLGCK664Iq4tPz8/Gd0SERFJL4kqLtS+Cc37/OvN+/zzUdOS288YfTkCtx3Y7pxbE5wvxw/oqoLpT4LX6uD6DmB8zPvHBW3dtY9L0J41Ro8e3fYzdOjQTm1//etfmT17NoMGDeLII4/k+uuvp6Wlpe39jz76KNOnTyc/P5/hw4fz0Y9+lKqqKh544AFuueUWNmzY0Daa98ADD3TZj9tuu43Ro0dTVFTEokWLaGhoiLv+0ksvcdppp1FSUsLgwYM58cQTWb16ddv16HTrhRdeiJm1nb/99tuce+65jB49msLCQmbNmsWKFSsOzx+eiIhIrGjFheg06YgpkFfkHw8cDKXHJK9vCfRZAOec2wVUmtnRQdMCYCPwOBDdSboY+ENw/DiwKNiNWgbsDaZanwFOM7NhweaF04Bngmv7zKws2H26KOZZyZUCOWOeeeYZPv3pT/PFL36RDRs2sGzZMpYvX863vvUtAHbt2sXFF1/M4sWL2bRpE3/961/57Gc/C8BFF13EV7/6VY4++mh27tzJzp07ueiiixJ+zu9+9ztuuOEGbrnlFtavX8/RRx/NHXfcEXdPfX09n/3sZ3n++edZu3YtH/7whznzzDOpq6sD/AAP4N5772Xnzp1t5w0NDSxcuJCVK1fy6quvcv755/OJT3yCzZs398mfmYiICOBPnz5wBrQ0wREfhq+/4wd4qcQ512c/wIeBCuA14H+AYcAI/N2nbwF/AoYH9xrwX8DbwOvAnJjnXAZsCX4ujWmfA7wRvOcngPXUp9mzZ7uubNy4sctrvRaJOLdsoXO3DPdfI5FDf2Yv/P73v3f+X6fvpJNOcrfeemvcPY899pgrLCx0nue5devWOcC9++67CZ930003uWOPPbbHz50/f7773Oc+F9e2YMECN3HixC7f43meGz16tPvlL3/Z1ga43//+9z1+3rx589x3vvOdHu9L5LD8/YqISGaLRJxbeopzNw32f24Z7lx9VdK6A1S4BPFMnxazd869EgRZHS1IcK8DvtDFc5YByxK0VwD9t9K+NxLljElCEdx169axdu1afvCDH7S1eZ7H/v372bVrFzNmzODjH/84xx13HKeddhof//jHueCCCygt/WBbpDdt2sTnPve5uLb58+ezZcuWtvPq6mq+/e1v89xzz1FVVUUkEmH//v1s27at22c3NjZyyy23sGLFCnbu3ElraysHDhxg+vTpH6iPIiIivVa/C3ZUtJ/nDoL8Ee3nnhe/Vi5J+jSAy0rRnSyVa5KaM8bzPG666SYuvPDCTtdKS0vJycnh2Wefpby8nGeffZb77ruP6667jr/85S/MmDHjsPZl8eLFVFVVceeddzJp0iQGDhzIggUL4tbjJfK1r32Np59+mttvv53JkydTUFDAokWLenyfiIjIBxYJQ/VGeKzDWFJrE+yv8wdjopUZor/jF69I2tSqArjDLdFOliSYNWsWmzdv5qijjuryHjNj/vz5zJ8/nxtvvJFjjz2W3/72t8yYMYO8vDwikZ63TE+dOpXy8nIuu+yytrby8vK4e1544QXuvvtuzjrrLACqqqridssCDBgwoNPnvfDCCyxatIjzzz8fgAMHDvD2228zZcqUHvslIiLSa5Ew/GAStNR3vjZ2dvtgTIrMsoECuL4R3cmSRDfeeCNnn302EydO5JOf/CS5ubm88cYbrF27lttuu43y8nL+9Kc/cfrppzNq1ChefvllKisrmTbN3yI9adIktm7dyvr165kwYQLFxcUMHDiw0+dcc801LFq0iOOPP56PfexjLF++nDVr1jB8+PC2e6ZMmcKvfvUr5s2bR2NjI9/4xjfIy8uLe86kSZNYtWoVH/3oRxk4cCDDhg1jypQpPPbYY5x77rkMGDCAW265hQMHDvTtH5yIiGSf2jcTB2+jpsOlz7QPxhSWwri5fvA2bm5SKzOk2JYKOVxOP/10nnzySZ577jnmzp3L3Llz+f73v8+ECRMAGDJkCH/72984++yzmTx5Ml/96lf59re/zWc+8xkAzj//fM4880wWLFhAaWkpDz/8cMLPueiii7j55pu5/vrrmTlzJq+//jrXXntt3D3Lli2joaGB2bNnc/HFF3PZZZd1qtTwwx/+kOeee47x48czc+ZMAO644w5GjhzJSSedxMKFCykrK+Okk046zH9SIiKS9UqPgbzi9vMBhf5r1WvxpbWi2SUMP6hLYrYJc0n88GSYM2eOq6ioSHht06ZNTJ06tZ97JP1Ff78iIpKQ58H9Z0LlaiidBjWbgJj46CubYMgYP0XYHVP9KdRQLly7qc9n3MxsnXOu04ZQjcCJiIhIdmuohu1B3YGajRDqsGRo/x4/yHMuZYrbaw2ciIiIZC/Pg+WXgvNi2mLWW+cVQ+nR7btPx831i9sXj0pqGhGNwImIiEj2iu4sTcjgixWwf3f77tNtQaaFJAZvoABOREREslG07GX+cMgLNi0MKIIx0eVmBhNP8Efa8kdAzqDoG+Enc/zUI0mkKdQOnHNYkqNqOfw8z+v5JhERyQ6xCXnHzPST9QJEDsAlvw5uMn+DgnNQ+3dobWh/f0s91GyG0ckrBqUALsagQYOoq6tjxIgRCuIyhHOO1tZWqqqqKCwsTHZ3REQkFcQm5N2+DvIKoKUhKJtVArlBeOR58MBZ/rRpbgGEm9qfUVCSnL4HFMDFGDduHNu3b6empibZXZHDKDc3lyFDhlBSktz/2EREJEXElr0cM9MP4sAP4h44Ay571j+v3gjbXvSPw00wchpUb4axs5KesF954ERERCT7RIvSF5TAstNg+0t+u+XAVX+Fp74BW1cDXSzBmXACLHmyz2uhKg+ciIiISFS07GUoBEuegryi9ms/Pxm2vkh78JYgXNoe1EJNEgVwIiIikt0O7IHW/f6xi/g/OLBQe4mt2FJbGIwvUyJfERERkX4TnT4tLPXzuRWWwoQyf7NCXqG/K3XcXDjzdlh6MuD5bf/3b5A/DPa/DyOnJjUXnAI4ERERyR6xKUTGz4PFK/xp1MUrgjVxI6Cprn10LbrZYfw8P2h78JzO700CBXAiIiKSPWJTiFQG69iia+GKRvoBXpRZe2BXWAqNNYnfmwRaAyciIiLZI5pCJFFB+ujo3B1T/fxvntce2EWnWlXMXkRERKSfdRxVi13H1tXoHLSvm1v8RPsUq4rZi4iIiPSTUMjP/9ZY45fKisof4Sf27TjCFjsy9+A5/ntVzF5ERESkHyWaKvU8eOgc2LHeD+IWPdEepDXV+jtUvbD/msT8b1EK4ERERCS7dJwqbaz2i9NXrvFzwO1YHx+kFYzw04uA/1owIjn9jqEATkRERLJL7GaEcXNh+WXwsxMhN9+/7jxYfmn7jtSmOj8PHPivTXXJ6XcMbWIQERGR7GIGix6H2jchfzj86Fh/5K2lPrjBtW9iKCjx18nF5oNL4u7TKAVwIiIikl08Dx76Fz8gGzfX/4lOn0aNmelPlUaT/o6bC/+2AYpHJX0DAyiAExERkWwTuwZu+1r4ygbA/GnTbeUwdhZc9mzn+0KhlAjeQAGciIiIZJvoGrjolGjRqMT54aL3bSv3R+Tyh0NDddJzwAGYi81/kgXmzJnjKioqkt0NERERSaaOBe27EgnD/Wf4O1Ojhe77sQ6qma1zzs3p2K5dqCIiIpJ9YktkdWf/bnjvZX99XPO++CoNSaQATkRERKQrhaX+BoY25p8neSeqAjgRERER8KdVG6rjy2uZwQX3QygnOA/BhQ8kfQ2cAjgRERGRROW1oopHwfgyP/HvhLL2AvdJpF2oIiIiIh3LazXVtgdqiXaoJplG4ERERERiy2slqrbQ200P/UQjcCIiIpJdEqUQiY6yNVYD3QRpvU0/0sc0AiciIiLZo7u1buAXtr9zWvy16OYGL9L9e/uRRuBEREQkezTV+pUVXMR/jV3rlmgdXEFJez3UMTP9nHCJ1sn1M43AiYiISPYoGOFXVAD/NX9Ye+qQROvgYoO6Hev9IK6rdXL9SCNwIiIikj2a6vxyWAAtjXD/Qn9ULVoeq6t6qNG6qYuf8J+R5DVwCuBEREQke8QGZF1NicZOiyZKIaI8cCIiIiL9KDYgi13f1t2UaDSFSApRACciIiLZJTYgW/Q41L4JpcekTI633tAmBhEREclOngcPngM/O7H7tCCRMFRtTGrakI40AiciIiLZqbEatr3oH2970T8vHh1/TyQMtx0Jzftg4GD4xjuQk/zwSSNwIiIikqWM9qoLsccxat/0gzfwX2vf7Ke+da9PAzgze9fMXjezV8ysImgbbmYrzeyt4HVY0G5mdreZbTGz18xsVsxzFgf3v2Vmi2PaZwfP3xK8N30mr0VERCS5ikbChPkQyoGJJyTeqFB6jD/yBv5r6THtlRmc69/+xuiPMcB/ds7Vxpx/E1jlnPu+mX0zOP93YCEwOfiZB9wDzDOz4cBNwBzAAevM7HHn3J7gniuANcAfgTOAp/rhO4mIiEi6M4MlT3Zf2zQU8qdNoxsdIH7n6uIV/j39LBlTqOcCDwbHDwLnxbQ/5HzlwFAzOwI4HVjpnNsdBG0rgTOCa4Odc+XOOQc8FPMsERERkZ5Fd6R2N4mXkwujpvn3NlRDZXl87rgk6OsAzgHPmtk6M7syaBvlnNsZHO8CRgXHY4HKmPduD9q6a9+eoL0TM7vSzCrMrKKmpuZQvo+IiIhkK8+D5Zf6Re0xGDc3aeW0+noK9UTn3A4zGwmsNLPNsRedc87M+nwC2Tm3FFgKMGfOnORNWIuIiEj6aqqF7Wv9YwvBhQ8kLXdcn47AOed2BK/VwGPAXKAqmP4keK0Obt8BjI95+7igrbv2cQnaRURERBKL3YDgeVC/C+qrerchIbbY/YSypFZn6LMRODMrBELOufrg+DTgVuBxYDHw/eD1D8FbHge+aGaP4G9i2Ouc22lmzwDfje5WDZ5znXNut5ntM7My/E0Mi4Af99X3ERERkTTnee0bEMbN9YO2ytWA+btRlzzZ/YaERHVRk6Qvp1BHAY8FmT1ygd845542s5eA35nZ5cBW4JPB/X8EzgS2AE3ApQBBoPYd4KXgvludc7uD46uBB4B8/N2n2oEqIiIiiTXV+sGbF4ZtQeAGgPM3JkSL2XcnReqi9lkA55z7BzAjQXsdsCBBuwO+0MWzlgHLErRXAMcdcmdFREQk8xWW+iNv217E32cZM22aVwQFI5LVsw9MlRhEREQkO5jBBff7GxA6am2CprrE70uBxL0dKYATERGR7FE8CsaXtZ/nFfuVGMbPS5wSJLpu7o6p3Re872cK4ERERCR7mMGiP8ARM2gLgxz+6FqiEbbYdXNJTNzbkQI4ERERyR6RMPzi47DzVcCDlnpwET+/W6LgLDZ1SHSULgWmVPujFqqIiIhI8nke3H8G7Hq187Wuqip0TB3iXNbWQhURERHpf021sGN953bL6b6qQmy91BSZUlUAJyIiItkhfwTkFfrHuQUwds4Hr6qQaEo1CTSFKiIiItlhfx20NPrHkWa4+Nd+SpEPUlUhRaoxaAROREREskN09MyCtCFFo9qnRj+I2CnVJFEAJyIiItkhumvU8IOvFErM+0EpgBMREZHs0FTrpwvxIv4GhMbqpKcDOVgK4ERERCQ75I+AAQX+cW4+/P7SlKuw0FsK4ERERCQ7xG5iaGmE7clPB3KwFMCJiIhIdigs9VOGRFOHjC9LejqQg6U0IiIiIpIdElVVSHI6kIOlETgRERHJHimQAuRwUAAnIiIi2cfz/JqmB7OJIQWK2SuAExERkexzsDVNDyXwO4wUwImIiEj2iI6eFZTAuLl+VYZxc3u/iSFFitlrE4OIiIhkh+joWeUaP2iD+KoMvVkXFy3HVblGxexFRERE+lxDNVSWt1diMNqPm2r9zQ09SZFi9grgREREJPN5Hiy/1A/YMH8ELhQ6uJG06E7WJFIAJyIiIpkvWgcVwELwyQf9oE154ERERERSVHTtWrQKQ9HItM4JpxE4ERERyXwpsnbtcFEAJyIiItkhBdauHS6aQhURERFJMwrgRERERNKMAjgRERGRNKMATkREROSDUDF7ERERkTSiYvYiIiIiaSZFitkrgBMRERHprdiEwCpmLyIiIpIGUiQhsAI4ERERkQ8iBRICawpVREREJM0ogBMRERFJMwrgRERERNKMAjgRERGRNKMATkRERCTNKIATERERSTMK4ERERETSjAI4ERERkTSjAE5EREQyj+dBQzU4l+ye9AkFcCIiIpJZPA8ePBvumAoPnOWfZ5g+D+DMLMfMXjazFcH5kWa2xsy2mNlvzSwvaB8YnG8Jrk+KecZ1Qfvfzez0mPYzgrYtZvbNvv4uIiIikgaaamFbOXhh/7Wp1m/PoFG5/hiBuwbYFHP+A+BO59xRwB7g8qD9cmBP0H5ncB9mNg24GDgWOAP4aRAU5gD/BSwEpgGXBPeKiIhINisYAXmF/nFeoX+eYaNyfRrAmdk44CzgF8G5AacAy4NbHgTOC47PDc4Jri8I7j8XeMQ51+ycewfYAswNfrY45/7hnGsBHgnuFRERkWzWVAetTf5xa5N/3lQLlWv8UbnKNe2jcmmqr0fgfgR8A4iGuSOA951z4eB8OzA2OB4LVAIE1/cG97e1d3hPV+2dmNmVZlZhZhU1NTWH+p1EREQklRWWwri5EMqB8fP888JS/ziU296WxnL76sFmdjZQ7ZxbZ2Yf66vP6Q3n3FJgKcCcOXPSf+JbREREuhZd4+aCY+cgFILFK/yRt8JSMEtqFw9VnwVwwEeAfzGzM4FBwGDgLmComeUGo2zjgB3B/TuA8cB2M8sFhgB1Me1Rse/pql1ERESyVVMtbF8LLuK/NtVC0Ug/iCsamezeHRZ9NoXqnLvOOTfOOTcJfxPCn51znwaeAy4IblsM/CE4fjw4J7j+Z+ecC9ovDnapHglMBtYCLwGTg12tecFnPN5X30dERETSRIZNlybSlyNwXfl34BEz+w/gZeC+oP0+4JdmtgXYjR+Q4ZzbYGa/AzYCYeALzrkIgJl9EXgGyAGWOec29Os3ERERkdRjllHTpYmYy4BcKB/EnDlzXEVFRbK7ISIiIn3J8zIigDOzdc65OR3bVYlBREREMkuG5XxLRAGciIiIZJbYnG/byqExSCGmSgwiIiIiKSqaBw7AebD8UoiEM2pUTgGciIiIZBYzuOB+P5Evzh+Nq31TlRhEREREUlrxKBhf1p5KpPSYztUZ0lgy0oiIiIiI9K2OqUQSVWdI492pGoETERGRzBStvGCWuDpDGlMAJyIiIpkvfwSMmakpVBEREZG0EM0Lt/2l9nNNoYqIiIikoGjet8Yaf+dp1I70n0LVCJyIiIhkHs/z871VrvF3n46bC5WrAfN3p2oKVURERCTFNFbDthf948rV8JWNQV44a9/YkMY0hSoiIiIZqEOA9rvFUDjSzw+X5sEbKIATERGRTON5fgmtAYXtbTsq/PVwGUJTqCIiIpI5ojtOt5X7Od/apPeu0440AiciIiKZo6nW37gQF7wB4+f7a98yhAI4ERERyRyFpX6iXsuJaQzBhff76USiJbXSnAI4ERERyRzRGqjXboKJH/GL2U8og/++HO6Y6qcW8bxk9/KQ9WoNnJnlAxOcc3/v4/6IiIiIHJpQyN9tGi1m7xzcOQ28sD+92lSb9tOpPY7Amdk5wCvA08H5h83s8b7umIiIiMghiRazLxrpT6uGcjOiDir0bgTuZmAu8L8AzrlXzOzIPuyTiIiIyOHhef6I2+InoKnOD94yYDdqb9bAtTrn9nZoy4wVgCIiIpK5oilF7pgKD5ydEWvfonoTwG0ws08BOWY22cx+DLzYx/0SEREROTTRlCJe2C+rdedUuP/MjAjkehPAfQk4FmgGfgPsBf6tLzslIiIictA8D+p3+a/j59EW7jjPD+Qa078iQ49r4JxzTcD1wY+IiIhI6vI8P1VItJD9uFX7G/gAACAASURBVDK48q+w9CT8FWBGpzqpaag3u1BXmtnQmPNhZvZM33ZLRERE5CA01fpltKK2l0NRCUyYD6EcmHhC2qcQgd7tQi1xzr0fPXHO7TGz9P/mIiIiknkKS2HsLL94PQAGFoIlT/rBXRbtQvXMbEL0xMwmol2oIiIikorM4LJnYdzxGTXi1lFvRuCuB14ws7/gTxqfBFzZp70SEREROVg5OX4QFx1x8yJw/xnw3sv+pobFK/wkv2msN5sYnjazWUBZ0PRvzrnavu2WiIiIyCGIVmGIhOEXC2DnK357ppfSMrNjgtdZwATgveBnQtAmIiIikro8zx95iwZvAGNmZnwprWvxp0p/mOCaA07pkx6JiIiIHIpo+SznYMf6+GuhAX57mm9k6DKAc85daWYh4Abn3N/6sU8iIiIiBydaPqtyDYyb669521YOBNUXtq/N7ClUAOecB/ykn/oiIiIicmhiy2dtXwsX3g/Xbgx2peb6AV0GTKH2ZgvGKjM73yzNxxpFREQk8xWW+kFabLD2+0thewUc8WFY9ETaT59C79KIXIW/Hi5sZgfwU4k459zgPu2ZiIiIyAdlBoseh9o3ofQYvyZq5Wr/2o4KaKiCIWOS28fDoDdpRIr7oyMiIiIih8zz4KF/8adRx8+D074Xf33/nowI4LpLIzLZzP5gZm+Y2W/MbGx/dkxERETkA4tdA7etHJ78Svu1vGIYOTV5fTuMulsDtwxYAZwPvAz8uF96JCIiInKwYtfAjZ0Fu14NLoTgixVpX4Ehqrsp1GLn3L3B8f8zs/Xd3CsiIiKSfLFr4EqOhofOaZ9OLR6V7N4dNt0FcIPMbCb+pgWA/Nhz55wCOhEREUktHdfALXoc9u/2R+YyYPdpVHcB3E7gjpjzXTHnqsQgIiIiqSd2DVzlGj94S/OkvYl0V4nhn/uzIyIiIiKHLLoGLjoClwFJexPpTR44ERERkfRgBotX+CNxGTZtGksBnIiIiGSWUKh92tTzoLEaML8tQwI6BXAiIiKSmTwPHjgLtr0IGEyYD0uezIhUIt0l8j3dzC5I0H6BmZ3a04PNbJCZrTWzV81sg5ndErQfaWZrzGyLmf3WzPKC9oHB+Zbg+qSYZ10XtP/dzE6PaT8jaNtiZt/8YF9dREREMlp0QwMADrav8dsyQHch6I3AXxK0/y9way+e3Qyc4pybAXwYOMPMyoAfAHc6544C9gCXB/dfDuwJ2u8M7sPMpgEXA8cCZwA/NbMcM8sB/gtYCEwDLgnuFREREWnf0ACAwfiyjNnU0N0U6kDnXE3HRudcrZkV9vRg55wDGoLTAcFPNP3Ip4L2B4GbgXuAc4NjgOXAT8zMgvZHnHPNwDtmtgWYG9y3xTn3DwAzeyS4d2NPfRMREZEsYOZPmWbgGrjuRuAGm1mnAM/MBgD5vXl4MFL2ClANrATeBt53zoWDW7YD0RqrY4FKgOD6XmBEbHuH93TVnqgfV5pZhZlV1NR0iklFREQkU4VCUDzar8KQIcEbdB/APQrcGzvaZmZFwM+Caz1yzkWccx8GxuGPmh1zCH09aM65pc65Oc65OaWlmTF0KiIiItmruwDuBqAK2Gpm68xsHfAOUBNc6zXn3PvAc8B8YGjMyN44YEdwvAMYDxBcHwLUxbZ3eE9X7SIiIpKtPA8aqsG5ZPekT3UXwP3ROfdN/CBpSfAzwTn3Tedca08PNrNSMxsaHOcDpwKb8AO56O7WxcAfguPHg3OC638O1tE9Dlwc7FI9EpgMrAVeAiYHu1rz8Dc6PN6rby0iIiKZJ5o25I5j/FfPS3aP+kx3mxhKAZxz+4HXD+LZRwAPBrtFQ8DvnHMrzGwj8IiZ/QfwMnBfcP99wC+DTQq78QMynHMbzOx3+JsTwsAXnHMRADP7IvAMkAMsc85tOIh+ioiISCZoqIZtqwEHW1/0zwePTnav+kR3AdwQM/tEVxedc92ug3POvQbMTND+D9p3kca2HwAu7OJZ/wn8Z4L2PwJ/7K4fIiIikiXM8BNe4L9m0KaFjroN4ICzgUTf3tHLjQwiIiIi/aKgBPKKoaXefy0oSXaP+kx3AdxW59xl/dYTERERkUOxvw5am/zj1ib/PFoTNcN0t4khc8cdRUREJPMUlsKEMgjl+q8ZUnUhke5G4D7bscHMSoC6YHeoiIiISOowg8Ur/HqnhaUZvQauuxG4IjP7XzN71MxmmtkbwBtAlZmd0U/9ExEREem9UCijSmZ1pbsRuJ8A38LfzPBnYKFzrtzMjgEeBp7uh/6JiIiISAfdjcDlOueedc79HtjlnCsHcM5t7p+uiYiIiEgi3QVwsemL93e4pjVwIiIiIknS3RTqDDPbh78bNT84Jjgf1Oc9ExEREZGEugzgnHM5/dkRERERkUPmeVm/C1VEREQkfWRRMXsFcCIiIpIZosXsvUh7MfsMpQBOREREMoNzZEsxewVwIiIikv48D5Zf2n6e4cXsFcCJiIhI+muqhe1r289bG/1i9hlKAZyIiIikv8JSGD8vODGYMD9ri9mLiIiIpAczWPIkNFYDlvH1UBXAiYiISGYIhaB4dLJ70S80hSoiIiKSZhTAiYiIiKQZBXAiIiIiaUYBnIiIiEiaUQAnIiIimcHz/PJZzvV8b5pTACciIiLpz/PgwbPhjqkZX8geFMCJiIhIJmiqhco14IX916baZPeoTymAExERkfSXPwLGzIRQrl+RIYOrMIACOBEREUl3ngcPnQM71vtB3KInMroKAyiAExERkXTXVAvbysFFYPu6jJ8+BQVwIiIiku4KRsCAguDEg98v0SYGERERkZTWVActDe3n217004lkMAVwIiIikt4KRkBufofGzM4FpwBORERE0ltDDYSb4tsss0OczP52IiIikvk6Vl4YOweKRianL/1EAZyIiIikt1CHcOaiXyuNiIiIiEhKKxoJE04Ay/Ffi0clu0d9LjfZHRARERE5JGaw5Ek//1thacaPvoECOBEREckEoVDGr3uLpSlUERERkTSjAE5EREQkzSiAExEREUkzCuBEREQk/XmeXz6rY064DKUATkRERNKb58GDZ8MdU+GBszK+kD0ogBMREZF011QLlWvAC/uvTbXJ7lGfUwAnIiIi6a2wFMbPg1Cu/1pYmuwe9TnlgRMREZH0ZgaLVyiRr4iIiEhaUSLfw8PMxpvZc2a20cw2mNk1QftwM1tpZm8Fr8OCdjOzu81si5m9ZmazYp61OLj/LTNbHNM+28xeD95zt1kWhNwiIiKS9fpyDVwY+KpzbhpQBnzBzKYB3wRWOecmA6uCc4CFwOTg50rgHvADPuAmYB4wF7gpGvQF91wR874z+vD7iIiIiKSEPgvgnHM7nXPrg+N6YBMwFjgXeDC47UHgvOD4XOAh5ysHhprZEcDpwErn3G7n3B5gJXBGcG2wc67cOeeAh2KeJSIiIpKx+mUXqplNAmYCa4BRzrmdwaVdwKjgeCxQGfO27UFbd+3bE7Qn+vwrzazCzCpqamoO6buIiIiIJFufB3BmVgT8N/Bvzrl9sdeCkbM+T5nsnFvqnJvjnJtTWpr5W4tFRESyQiQMVRv91yyqwgB9vAvVzAbgB2+/ds49GjRXmdkRzrmdwTRoddC+Axgf8/ZxQdsO4GMd2v83aB+X4H4RkYzkeY66xhZKivLQni3JepEw3HYkNO8DywEDxpf56URCmZ/mti93oRpwH7DJOXdHzKXHgehO0sXAH2LaFwW7UcuAvcFU6zPAaWY2LNi8cBrwTHBtn5mVBZ+1KOZZIiIZxfMcFy9dTdl3/8RFP1+N52XPSINIQlUb/eANwEXAi8DWF/2RuCzQlyNwHwE+C7xuZq8Ebd8Cvg/8zswuB7YCnwyu/RE4E9gCNAGXAjjndpvZd4CXgvtudc7tDo6vBh4A8oGngh8RkYxT09DM2nf3ALD23T3UNDQzavCgJPdKJEk8D564JsEFlzXTqH0WwDnnXsAf0ExkQYL7HfCFLp61DFiWoL0COO4QuikikhY6/p+pJlAlq+17D3auT3YvkirzJ4lFRDLAiMI8igf6/+YuHpTLiMK8JPdIJEnCLfCjY7u+ngXr30ABnIhIWtjd1EpTSxiApuYwu5tak9wjkSTZVt71tbFzsqaclgI4EZE0UFKUx5xJw8kNGXMmDaekSCNwkqVGTO762oUPZUUhe1AxexGRtGBmPHxFmdKIiAwe7acLqUwwErd8MVz2bFZMo2b+NxQRyRChkFFaPFDBm2Q3LwKt+xNfe+9laKrt3/4kiQI4ERERSQ+eB8tOh12vtrflFsDY4yGUC+PnQWF2VFzSFKqISJpQJQbJeg3VsKMivi3SDBf/CizkB29Z8t+GAjgRkTTgV2IoZ922PcyeOIxHrigjFMqOX1QibToGZ5YDE8qgaFTWBG5RmkIVEUkDNfXNrH13NxHPsfad3dTUNye7SyL9r2gkTDgBCMGAIr/qgudlTfWFWArgRETSQMfBBee85HREJJnMYPETsOgP0NoAeFC5Ghqzo/5pLAVwIiJpYERhHtPHFredn/qj5zlwIMzmnfvwPAVzkiU8Dx76F3jo3JhGIxuLy2kNnIhIiguHPS5cuprXdtS3tdUfCDPru3+iqSVC8aBcXr7hVHJz9W9yyXBNtVC5Bgj+0dK2Bi47qi/E0n/tIiIpzPMcn1y6mpe3vR/XXpAXoqklAvjB3JaahmR0T6R/5Y+AMTMhlOOvhbt2Eyx5Mus2MIACOBGRlFbX2ML6DsHbE1efwGs3nkrxoPbi9lNGFSWjeyL9x/PgwbNh+0t+Ml/nsiptSEeaQhURSWElRXkcM7qIzbvaR9j2NYfZsz/C+us/ztu1jUwZVUQoC0oHSZZrmz4NVJb7bVk4fQoagRMRSWlmxhNf+AjRlG8h4K5VbzH/e6v4zLK1TBlVrOBNskNhKYyZ035uIcgflrz+JJn+qxcRSXE5OTnMGj8EAz40tph1294n7Dkq3t3Nm1X1uCzMgSVZyItAS/tGHlwEat9KXn+STAGciEiKq2ts4ZXt+3DAG+/VM33cEHIMCgbmctaPX+DipeV4noI4yWCeB/edBjUb29vyiqD0mOT1KckUwImIpLiSojxmTxxGbsiYM2k4y6+azxNfOpHG5jCRYCSurrEl2d0U6TsN1fDeuvi2z5dDFi8f0CYGEZEUZ2Y8fEVZWyF75+DG/3md6KBbQV4uwwsGJLeTIn0p0U7T5n39348Ukr2hq4hIGgmFjNLigZgZVfsOULFtb9u1ptYIu5tak9g7kT5WNBLGzms/HzgYRk5NXn9SgAI4EZE0Eg57XP7A2ri26WOHUFKUl6QeifQD5yAnBzA44sPw9XeyevoUFMCJiKQNz3Nc+PMX2bgrvurCzz4zC8vSZKaSJZpqYftawEHVG3Bgd7J7lHQK4ERE0kRdYwuvVu6Na5sxdjBmplQiktkKS2H8PAjl+q+FpcnuUdJpE4OISJooKcpj5vghrAuCuEG5YCFj/vf/zOyJw3jkijJCIY3ESQYyg0WPQ+2bfuoQjThrBE5EJF1EIo7NVe3TpwfC8ErlXiKeY+07u6mpb05i70T6UCQM958BPz/Rr4fqecnuUdJpBE5EJMV5nqOusYXa+gM0tkS6vE+DEpKRImFY+s9Q9Zp/vvVFPy/c4NHJ7VeSKYATEUlhnue45N5y1m3dw8zxQyjKy6EhCOJij+dMGEJp8cBkdlXk8PM8WHZ6e/AGgNO/VtAUqohISqtrbGHd1j2EPcfLlXt5+KqytsL2DS0RCvJyAPh7VSORiDYySIZpqoUdFfFtAwqhoCQ5/UkhCuBERFJYtIxWjsGHxg7m1sc3EFv2tCkYgatvDrOlpqGLp4ikqfwRUNIhYW/rfthfl5z+pBAFcCIiKczM+PXl85g+fiivbt/LS1vfT3hfCDiqtLB/OyfSlzzP37BQuym+fexspRFBAZyISMrbs7+V17fvjRt568gD6lROSzJJUy1sezG+beRxcPmzWgOHAjgRkZQXnUbNDRlzJw3jqS+fSKJfX/qVllye56ipb1ZS5cOlsBRGTIlvO/kbWV9CK0q7UEVEUpyZ8fAVZdQ1trTVPJ175HAq3t1NQV4ujS1hZowfqnqoSRBN8TK8YACf+sUa1m3dw+yJw3hYSZW75nn+6FphafcjaWZw1d/guzHTpcec1ff9SxMK4EREUlw0SCgpymureRoN6IYOyuWT95bz2va9XHLvGgUO/Sia4qXi3d1MGzOYjTvriXiOiq17qGtsUVqXRKLr2irX+CWxFq/ofkQtLw+ur4K3VsKUhZCrsCVK45AiIiksGiTM/94qLl5ajhcshAuFjNLigbx/IMzr2/1qDOuCwEH6R11jCxXv7ibi4PUd+9r+bvIHhBheMCDJvUtRTbWwrRy8sP/aVNvzewYMgmnnKHjrQAGciEgKi80DlyhAi10fN3viME2j9qOSojymjxvSdh5d+dbQHKE2+Hs6cCDMU6/tJBwOJ6GHKcbzINwMucHIZG4+eBHobs2g5/lVF7SusBOFsyIiKSwaoEXXVnUM0KJpRt6srmdEoYK3/mRm/PRTs5j/g+c6X8MP3o65+Zm2ts03n86gQVn6azfc4pfDqn6jva21Ae481p9KXfJk56nUaP3T99bD+LKep1uzTJb+L0lEJD1EA7QtNQ1MGVXUtgYuyvMcn/pFOWvf3QPAh8cN5r8//xFyckJx93RcQyeHzvMc1/z2lU7thXk5jCjM49kNVXHtf/57NWfOGNNf3UsdkTB8dwJ4+ztfcxE/VUj9LhgS82fjefCLU2Hnev9c9U87USgrIpLC/ABtDWf9+AUuvndN2zqrqLrGlrbgDeCV7fs4/54X8TyH5zmq9h3g4qWrKfveKi6KWUMnh66usYX12zonVt7fGmF3UyunTotPNrvsxXey88+/amPi4C1WQ038+b732oM3AJymUTtQACciksJq6ptZ++5uIp5j7Tu7qalvjrs+vGAA+bnxo2qvbN9LVf0Bf/PDd1ex9t09Xb5fDl50ervjmKYBQwflEgrlcNzo9uoYFVvfp2rfgX7tY9JFwnDf6T3fFzsy7Hnw64v6rk8ZQgGciEgK8zwv7rzjDGhtYwv7w51HJt4Pdkh6HdpbI5HD3MPsZWb88tK5TBlZENcecfDS1t3U1B/gjV2Ncdfeb8qyXcJVGyHS1PN9+UPhH3+FvTuhsQZq3uh8j9a/xdEaOBGRFBUOe5z6o7+2nR8/cWin3GKJVrQVD8zlyOEF5OUa+1vjg7utdY2MG66aqYeD5zk+uXQ1f6/uHKB86hdrKczLiWvLzzUmj9SfPSOOgX2V0BoEt6Nnwl0zIPrPjcKxnd8zfj4Ujey3LqYDhbMiIilqS00DDc3tI2a3nntcp00IIwrzKOgwhdrYEub8pas7BW8Ad63akp3rsPpATX0zr2zf2+X1xpb40c79YcfM76yiqamV1VtqiWTDaGjHoGtAAVz9InxzG1z1PIyZA7tehtix4sYd8e/5zONw2VOqf9qBRuBERFLUUaWF5Jg/JZdjMHlkUad7dje10tRhCnXKqCI2vFef8Jnrt72vKgGHycHEE/XNYY77zrNEY+jHPz+f48YPJdTX04OeB43VgPlBVX8FQ8Wj/NGzytUwegZc8RzkBCOTRSPhvYqenzFxvoK3BPrsfzFmtszMqs3sjZi24Wa20szeCl6HBe1mZneb2RYze83MZsW8Z3Fw/1tmtjimfbaZvR68527T3ngRyTB79ofbftF7zj/vqKQoj+MnDm07L8gL8fjVJ1CQF/9/7yGDHCX7PaxGFOZRMKD9V0/HzSRdiR0A/Zd7VjP1hqdobm493N2L+UAP7j8Tfng0/HAK/OLjcKijf71NsGsGl/4RvvYWXPWX9uANer+rdPeWg+9nBuvLkP8B4IwObd8EVjnnJgOrgnOAhcDk4OdK4B7wAz7gJmAeMBe4KRr0BfdcEfO+jp8lIpLWhhcMoChI/Fo0KDdheSYz47dXncDq607hqS+fyBs3n87eZo/9LfHbFzwHD112PI9cWaZccIfJ7qZWDsRMUyfaTNIbzR4cfdOztLT00ZRqQ7U/Aha1owKWne4HYQcjWs/0jqnwwFk9PycUOvhRv4GDofSYg+tnhuuzAM4591dgd4fmc4EHg+MHgfNi2h9yvnJgqJkdAZwOrHTO7XbO7QFWAmcE1wY758qdcw54KOZZIiIZYXdTK03N/qhbU3OY3U2JR2lCIeOIIflMHlnMm1UNDB2U0xb4xTqqtHMi4I7CYY/NO/d12v0qnQ0vGMD08e2ltI6fNIzpY4sP+nmr3+lFXdCDkWik6711vatDGhUJB/ncPP99lWv8eqZbV0P1pq5H07obqesu3j3rR3DF8/DvW7X7tAv9vQZulHNuZ3C8CxgVHI8FKmPu2x60dde+PUF7QmZ2Jf7IHhMmTDiE7ouI9J+SojxmTRjGum17mNXD1Gc47DHzP1ZSfyBM4cAcmprjR3OOnziUkYMHdft5sc8oHpTLyzecSm6ufnkmEk2w/Pr2vcwcP4SffWY2IwcPYtf7+xOW1uqNnL76o44kSF2SOwjyh/fy/WH4wSRoqYfQILj4YRg7ByrLAQ9+doJf6urCB/01b9F/JHieP0JXuSZxuaxEgZnlwIQymLNE6956kLT/MoORs37ZCuWcW+qcm+Ocm1NaWtrzG0REUkBrq8fGnfvwHGzeWU8k0vX/ZW6paaD+gD9a19gcYdCA+P97b26NsKmHkbU3q+vbnlF/IMyWmobD8C0yU11jCxVb9xBx8NqOfYRCIZyDLz38cqd7jx1TzPNfP7nHZ9696u24HcItLZFD360aboGfzO3c3toEy07r3TRqzWY/eAPwDsBv/jUI3mJUlvtTqvef6Qd8DdVQv9MvkxVbLitW0Uh/g0PU2Hlw7SY/0FPw1qP+HoGrMrMjnHM7g2nQ6qB9BzA+5r5xQdsO4GMd2v83aB+X4H4RkYzgeY7zf/5iWyqK+mY/oDrmiMEJ758yqojiQbldjsC99l49Z979AoV5Obzy7VMZMCA+R1k47PHvy+Preh45vPsRu2w2vGAABQNyqG8OUzAgh+EFA6ipb6YiQWmt/Nwcxg4r4n8+P5/z7mlfi2bA/3xhPv/609V4Ln6HcEtLhKk3PU00Zn/+qyczrqTnKfA4ngdLT/GDrkR2rPenQ3vKr1ZQ0tsP9AO1ZafDjnUw7J/iLzfUxNc7jW5wSMbu2AzQ3yNwjwPRnaSLgT/EtC8KdqOWAXuDqdZngNPMbFiweeE04Jng2j4zKwt2ny6KeZaISNqra2xhw459beeFA3OYMqpzGpGoUCjEyzecyh+/fCKrvnIyxx85PGGS38aWCP/6sxfjRno8z3HBz17k9ffiR9ym3ryy7xbWp7ma+mbqg/WJ9c1hauqbu4w91lf6gVleh+noY8cM5kNjh3L8pOHkdtghvG7bnrbgLUSYy3/4az750799sBx++96D6te7vj6gwB8t62k3aGEp5BR0f0+sHRWAgz0dd48m+AMKhaB4dPzUq/RKX6YReRhYDRxtZtvN7HLg+8CpZvYW8PHgHOCPwD+ALcC9wNUAzrndwHeAl4KfW4M2gnt+EbznbeCpvvouIiL9raQoj+OPHE4I+NDYwbx246k95goLhYxbntjIR37wHK3hCKu/+c8cP2lYp/s27NhHTX37qExXCWk95wcS0tmeDiWx9jT5I2dzE/x5zwkCs2OOGExRUJ2hIC/EH75wAqFQiIevKGP1dQvidgjPCjZHhAjzysCreGbgv3Nf9QXU1feiLBX4gdmWVZ3bz1tKWyDV2gh3ToVlC7ufSq3f2btyWD0pViWFw6nPplCdc5d0cWlBgnsd8IUunrMMWJagvQI47lD6KCKSqsyMX18+jy01DUwZVdSrRK91jS289O5uPAcvV+7lc79cx2P/9wTqmlo45f8915bw1wGn/PAvbVOp3Q18zByXeMo2G4XDHm9W1zOiMI9Jw/Ljrv1TSSFmxiNXzqeusYXBeTlUbNvDUSMLGTk4HzPDzHjlxtM6/Z2GQtYpsfI7u/cDcBQ7KWY/ZlDMfoob3oYh07vvaOsBuO1If51bR0ee7Jeu2rW+va1ytT+NWTw68bN+dGzPfzg9GfZ//FE2OWy0vUhEJAV5nuPT963h7B+/wCX3runV1NnwggHkx9TffGPHPi5cuhrDOuUoa2yJ8K/3+FNyQwZ2/W/5C3v52ZkuHPaYccsznHn3C8z73p+Z9d0/xV3/R50fLEWDsYEDc/nI5FJGDSmIW7eWmxvimCMGxwXknueoqW/GOdeWxuX/jMhnUC5sYRSO9llOu/dkaG3uuqORMHx/UuLgbcJ8KBweH7y1SRDFex78/GNdfxbAyOPg2s1weYLRvlifX6Mp0sNMpbRERFJQXWML67buIew51m3d06vyV7ubWtnfYfPCK5V7iTiPwoE5cXVVAd54r56ahmb+Ud31btM3duyjtqG5xxQkmW5z1T4aW9unGWPrzBbmdb8+sTue57h46WrWbd3DzPFD2LyrgYaWCCHzg7bTS/ZhDbHhlYNfnOLXEU00Klv7JkT2J/ikkJ/mo/KlzpdGz0i8kaGpFmo3df8FPr0cBh/hj95FS2bZAHCtUDwJTvwKzPo0DOichFoOjQI4EZEU5OeAG8q6re8za8LQXpW/KinKY/bEoby0NX4n5N6mVva3Jl7j5Jxj1vghGO15nSaPyOOtOn+NlwM+/8sKfv/5jxAK9ZwE+I2d7/PengOcNm0kubmZ8SsmHPbYVtvY5fVV15580LVMaxqaWfuuv86wYlv7OsTooOfK3SNweQMxYkbdqjZ2vXu09BjIye8cxB3xIf/+RDtKz/lJ4tGxwtL2oKwjy/Fzuw0Opl2jO0qbaqFgBDTV+e/XqFuf0RSqiEgKam31eGPHXiLOsfG9fd3mgItyDlo73Ddz/FCOHl3M7AlDHKiCHQAAHnFJREFUE74n7HnM/I8/tQVv08cUcc3p0+LuWVe5l117E43qxDwn7PGhm5/ivP9azdW/eZmjbniGAwc6125NN9Hkxlc//ErC60UDcygtPvjRSdfDDlDnjMjlT8c35g70g6REvDBc+ACdfr0v+I4fTOXkwLeqIBQzmvvMvyfexBANyq7dBFc+D2PmBIHbfL/t0j/GB2jRklmhHKUE6QcK4EREUoznOc776Qs0BaNmDS0R3qyu7/F9NQ2dd5M+fPlcQqEQd188s9P9U0cVcuWDL7V9DsCGnY1MGt45ZcTbNd1//paaBvZ3iNeee7Omxz6nutjkxok0NEeobUhQ6aCXegpxPAdv2ZGQG/N3Et4PDVWdbw63wH+UwiMXAbEBWQiOPLH9NG8QXPNK+6dvWw2NXfxdhUIweAyMmQ6fWwlf3QyXPaW0HylAAZyISIp5b3cTG3fFr0sbUdjzFGqiX6cvJ0gPErWpqpGNu+KnBqceUcy0MUM4bnRhXPv3n9zAb8rfZeeehoSjRkeVFnZqO3Vaele+8Tz3/9u78zA5qzLv49+7lk7vS5JOpzs7CVkgkISErIBoJMZlYHTgHZh5AXEJLjg46qW+KiAKjsu4zSAoLy64AY6g4ga4REBDgARCyGYgJOns3Vk6vVZXV9WZP57qLV29ppauzu9zXW1XPVX19HkOZdWd+5xzH275xeZ+n9dfFq0vAynMO6YoF244JQt3dFfP+m3Vp+yOALDqP+CWo17mravCcjrfMQ7+5/rEWbhIGHb/FaLR09uUXpJOAZyIyDCz+UD3oGv62PwBLSIoLxrFwi6bq/sNFk/z6pL1N3+t3X3XLcTn8/GrD13C7IrOrM/WIy186pdbWfalJ7ny7nU9VqbW1HdfGTmzPJ/+80vD27GmMC/uq+/3eQPt20TKi0axaHJJr4/Pn1js/bcff643fNnuh2+Du5dDuLVzs/hTdz4wHyy5sWfwBnD0Fbpl6fY913Nz+7YQ3FkB97/Vy+xFhp5plORTACciMkzEYo4Ddc3kBrsHR7/70IoBZWrMjIduXM6jNy3nJ+9ZzM47VuOPf3mXF41iwaSegUJeoPO8i6eWUVHi1Tfz+318950J9tAENu6ro7axM2CLRGKs/Ppfuj1nZ20zc297jEgk1vGcbQdPcuRky2llrNKpfSFJX+ZNLO53dXBfzIyfvW8F6z75euZWFeMzKMjp/GrOCfi9RFvL8Z4vrt0GXxgHX53tbZm195TFBjdvTRy8gbfYIaeo8/6kJd6ig1jM28Hh4Etwz0Xg4kGei8KedUO+Tkm+kbFESEQky8Vijqvu+SsbE2R86sOO8gHECO214zbuPcHCKWUse+/SjsfMjJ+/bzlXfucZNlXXMWd8AT6/n60H6plbVcR91y2ioiSvW6B4vKn3jEvXcHLH4fpuZTXahaKw7fBJZo8rZv7nn+iYa7d4ahkPrll2WpmrdPDizO7XlR80mrtc673XLhrc/qQJ+HxGVWk+j950EceawkQiUZZ9aS0Az+05QW1DKxXF5VA5Hw5uTNDQKBx6AX757s5jOUV9F871+eATe7yN6vPHes91ztuVYX+CoVjofeGEZIQycCIiw0BNfShh8LZggCVEIHHtuK78fh8Pv2856z+1kpxAgC0H6nF49eDe95MXekyp8vdSGmPx1LJuWaeTLb0Heidb2nj7PX/ttlBiw56ebRuOjjWF2VjdfTj7ZzcuY8GkUvw+Y/G00Umtj9deBNjv797vZvH/ectXB36ycBO0HOv7Of4AjJ/rlQIx83Zj6C14C+R7w7gybCgDJyIyDBxr6lldf3ZFAQ+/b9mAMzxeHbiyjgxcosDP5zOikViP1aqb9p2ktrGVii4ByezxReQHfd2Cr7uumc+8ScU8s+soS6aNxjmjNC9xkVYfcHZ5AVsPdV+QkZvjZ3R+kFjMcawpzNjCHKJRN6htw9JhdH6Q/BwfTWHv+otGBTinqoSH37+8o92nm31LpH1P1fb/jh3Bsq+X4dBEKs/zhkQHKhKG6ud6f/xdjycuHCwZowBORGQYGFvYc4y0MDfIYBYCmBkPvHdpv8HFa8cS77wQi3Zfhejz+Xjp1lW8/Z6/seVgAwbcdEo9tPwcP83hKHl++PL/mc+FU0opGZXD27/zDDuPNLDmhxt6/J3WthhHm8J88Mcb2VBdx4QiH0dboDUSoyg3wIufuQyfz1IaJA3E8eY2QvEtyAz4Y5eCvacz760/XfdU7Xb9iQr39ubqhwa+WjQShjsq6F565BQFCQoAS0YpnBYRGQbGFecyb0L3jeM37K2jtqGPfS+HaEZ54m2f6lraehwLBv08etPF/OTdi0m09KA57G3P1RKFDz+4iZsf2kx9OMqOww3EHGw+2DNYXDCpmHBbhA3V3o4RBxpitMYXOzSEInz19y+z+ptPcuGdf+T1X1lLa2tmhlvHFuawcHIZfp9xYZKHS/vTPpzaLXgtqvCK6OKDiUvh5pch2L5SOADlc7yVqpOXd+6QMBDV6+kzeIPBZf8kLZSBExEZBqJRlzDZNpjkUyzmuOb/r+8YenvgvUsTLhSoKMljwYRCXjzQGVwV5viZNb6ox3PBCyZmjOt/r8+ogw17T9Aa7XsHhg3VJ7nirqd6ffyev+3vuL3neAuzbvsD225dRX5+evfTdM6r8db5k+ESaF23qyoo94rvRuPBrQ+47pde6ZDBbmFVMrnvxycvH1z2T9JCAZyISIaFw1Eu/9bT7DjSWVTXgEVTSgc1VJdoEUOi15sZD3/wEg7Vt7CrpoGxhaOYPb64z7ln44pzmVORz/YjzX22Yc74Irbu77922vGWwZUS+dXmg1yzdMqgXnO6ahtaeX6vt0/p8+2rQUvSl4VLqL2YLsT3Kl0C+571fhcOYXeESBj+e37P41f/HKrmegGhivcOSwrgREQyKByOMu9zj9MS6Qxo2jeW3364gWjUEQgkbxFDO5/PmFCaz4TSnttmJWJmfO+dSzrKW/Tmk6tnsnjq6AGdczAunZX+EhaRSLTb/WEXw5jB9b/pzMgNpYF713NqqRQAqs4b3DCspJ0COBGRDAmHo6z6+tpuwRt0fp02tkZ5tbaR2ZXFPV+cwEAXMQzV+NI85k0o4qUD3fdFbQ84fcCy6WM53txzCPW8ygL8PmPTgZ5z4uZNKOalA17WbkwuHAt1PuYzWDiljPGlPbfqSqVwOMpFX/lLx/1FkweXDU2brhm5wYrF4M939DweLIxvtSXDmQI4EZEMiERinHf7Y7RGez7WvrKzKDfAzIr+55511T75PRXMjEc+cBH/cNfTbOtSGuTpT1zKvmMtLJ5Wht/vZ2yhj8VTy3hujzf8OH9iMY98YAVg7Ktr4k9bDxFqbWPy2EJWz63C5/NT29iK4a3ubGuLsbH6BIumlHKwIcT6V44RiUQIBtMzBy4SifGzF6q75aU+tHJGxlbDpkxTDRx4tufxaAs0H9O8t2FOAZyISAa8WtuYMHgDeHDNEnL8/mFVE62d3+/j1zddzJXffobNB06ycEoZE0rzmVjWmSFrL4PRNShrD36mjC7kXRef3eO8XevP5eT4WTZjLM3Nbbzuy08C8IlfbEnpQoZIJMartY1UFeWw+ItrCUW6r8r0j7DYzZPoogwmLR1cDTnJCAVwIiIZMLOikFy/t91UVwU5fuZWlQy7wK0rv9/XbzFbn8+6BWVD8euXD/W4/89L+lkxOQThcJQFd/yBpnAvETXe0PCIkz+azgHwuBuf9nZnGGnZxhFo+H5CiIiMALGYo7ahtccG7j6fj5c/u5o5FZ2Zq7lVRWy+7bJhHby1S1inLMnecUFln/eTIRZzXPGtp/oM3u5/50L8vW0Kn82OvkKPBQxacZo1lIETEUmRSCTGVfc+w+b9J1mUoC6bc/CZt53LWeX5+H3+lAdE2SYYDLLt1lX8avNBLp01hkAg+V9ZRxpC/ZZGSbRLxogwdiZeHic+XDx5mea9ZREFcCIiKRCLOa76zjpe3OftOfrs7uPUNoSoKMkDIBSKcM7tjxNz3vyq7bevVvCWQG5ugF++dJBbHt3aZ3HioTrR1PcuD/kBY05VSdL+3rDSctxbxRqLeTs4XHW/sm9ZRAGciEgSxWKO2sZWDp1s7gje2q382lNsuuUyYjHHm/7rKWLx0auog43VJ1g2YwTOszpNx5rCPLf7OA54bvfxXosTD9WYgu618ioKgzy45kL21YUHVOA4q/UoBKzsWzZRACcikiSxmOPqe5/pKJ9xqsbWCNsOn+Squ9fR2mWRo89g8bSyNLUyuxTn+DtmaTkgFArjkljjblxxLosml7Khuo55E4v5xQdW4PP5mHYmxDLJKAQsGaMATkQkSY41hXsN3tp95MFN3YI3gO9dd8HInCSfBBv2de/Pi7/6FIumlPLge5dR29RKXVOY6WML2FnbiM+M2eOLEmbMYjHXbdVsKBTht1sPcrIlxInmVnwGOQE/iUtrjGCnUwhYMkoBnIhIkpTm9v+R+urRnhPmy4syvL/mMDZ9bM8dGDbsreMdd/+VzQcbejxWNCrAi7dcRiDQGcSFQhGuuPuv/L2mifGFftZcOpXP/WZXj9cOm/1ORQZghA7si4ik366jTf0/6RS5fkbuJPkkqCjJ49yKvB7HEwVvAA2tEZ7ZVYNzjkgkxgvVx5j92cf5e4333+ZwYzRh8NZOo4iSLZSBExFJkqllPQMNgNuvmM1tv9rR4/jk0Xms/ejrRu4k+SRwDgKBINAy4Ndc+/2NjMuFE2Foi/X//HbzJxYPz/1ORRLQp4aISBJEIjEuuOMP3Y75fcaSaaN505zxCV/z0HuXaO5bP441hXk5vtH9YNSEBhe8nVdVyCMfWKFSLpI1lIETEUmC7YfraY50r2r/m5tWMLuymFjMkReAlkjnY4smlzC+ND/Nrcw+YwtzWDS1rN/FIQPVXra2qijIQzcuIRjMwWep31VCJNkUwImI9CESibHjSH2fKxwBTjS1drs/a1wBsyuLMTOON4fpulPTbz+0gnOqShQwDICZ8eCaZdQ2tnKkronL717f7fGgwQ/efSHzK4s57/N/oq+k27wJRTz8/hWcaIn0uoerSLZQACci0otIJMbc237fseF8DvDdGxaxfMbYjqHP9nIUX3tse7fX/uCGCzsCBC+LNJqNe0+wcEqZgrdB8vmMiuJcxhWNYsaYPF491jkf7ns3LGLFDK8Mxs473szLB0/w0Z9t5rWjzR3btOf6Ye3HXsf40gLMjPIiDVtL9rNTN1ge6RYtWuQ2bNiQ6WaIyDDWXjNs//FG3n7P+oTP+fgbz2Lx9NFc+Z2enydzKwv59b9d0i1IO7UOmQxNW1uEWbd2bkG2847VPeYRtvd1aW6AXUebmFlRqIUikrXMbKNzblGP4wrgREQ6ebsprOe5PceHfI67r5nPW+ZNSGKrpKtwOMrG6hMsnlamRSAy4vUWwGkIVUSki9qG1tMK3gBWnVuRpNZIIjk5fu0bK2c85ZRFRLo4nVGJALDzc5cRCOjfxiKSWgrgRETiIpEYWw/1Xq5i+phR5AUTf2zOGlfA3+9cTU5OTqqaJyLSQf9MFBHBC97Ov/0xmtt6ZuAumFjEt69bTHnRKKJRxys1Ddz8wEZ21nauhrz1bedoPpaIpI0COBE540UiMX7+4t6Ewdu6j19CZVlhx8rRQMCYU1XCb//tEmbe8jgObyhj6fQx6W20iJzRFMCJyIgUicTYfPAE2w7Uc96EIoKBYEch3vbivKFIhE17j/OF372SsADsD6+/gKrRRQnPHwwG+PvnVms1pIhkhAK4DAqFIvxhxxHOKs9nTmWJ6hSJDEBvJSRCoQhrd9Zy8fQyfr31EJ96ZBun5tMCwDeumcu/P7Sl330yDVgxc1yfz9FqSBHJFAVwGdLYFGbu5zs3vvYDD9+4mPOnjAGMfXVN/GnrIVrDEWaML6SlDSaU5XLkZBurzhmnVW5yRonFHPvqmnhi837ufGxXx/Ef37CAzfvq8fna+OIT1f2eJwLc9MCWfp93/3ULuGhWhbJqIjJsqZBvBoRCEc67/XESTLfBB1QU+TnUEO35YBcvf/oNFBXlJaWgpSrESyLtQdNTO2qYMiaPSWPyWffKcc6pKuRocxtH6kIsmFTKnKrSIWeP24c5X9x7nFAozKhRPlpbY91+h0JRfvL8/n7/P5Es2297A3l5eWn5WyIi/VEh32HCW+mWOHgDiMGAvqjOu/PPfPzSSXzlL/s6hok+9cazaHMxCvKDzCgvZtn0cqJReHpXDbX1rfzTBRMIBoMd7dh04DjP7qzlp89Xc6A+QkW+ce3SyeTm+nt8iSb6nZcXoCg3SEMogt+MSaPzGRUMsnTaGGUuslhjU5gfrtvFj57bO+Cg6aOvn0Je3sDeN7m5fnDQ3NLG19dW97n5eLpUFvj49D+cy+q5Vcpui0hWUAYujcLhKD9Yv4sv/O6VjPx9gJsvnkxBgY8vPLYnpX/nI6+bQn5B31/o7V/kodZor88pKRzFytnjOzah7k+qttiJRGK8WtvIzIpCYjE6bg818xQKRfj1lgPU1DUxKscGFPik43dDQ4hvPn0waf2WSRX5Pr501TzOGlfIu76/gVePtnR7/D//aQ7jSvKZVVHMuOI8ZZ9FZFgasXuhmtlq4Jt408juc859sa/npzqAC4ejPL2rhj1HGwiF2jq+HAMBx388vnfQ5/PBsMhQZFp5Lly/fEqf2cHm5ja+9pd9Ha8ZSBA5kN8tLRH++6lqwvFkVNBHxwT4T686i3B0cOdragrz9Sf3Z6AXR4aq4hwO1oe7Hfv0m84iho/XzxnLkztO8NbzK7qV/ojFHEcaQtTUt9DYGmXptNHKEotIVhiRAZyZ+YGdwGXAfuB54Brn3LbeXpPKAC4cjjLz1scG9Zqtt76BPcdDvPWudd2O33H5bP75wsmY+dlxpJ5ILMbJljaCFuFfvrspmc0WGRb6GobNzfWTH8zhyoWT8PsDCsZE5IwxUufALQZedc69BmBmDwJXAL0GcKn0fPXgNsBunyx9bn4eOz+3mmd2HyXgtx5zyOZOKO32ui23lPOjZ14jmOOIRIz5U0p4YfdJcnJcjyzUQHxm9XR8fsM51+dw5qlZqW+srSY908olEz6zahoF+blcfn4Vz1XX0dQa4fyJxbx2tJmg38f8qhIe236EuuYWQqHIgLOQXYfO8/ICVBTl86ZzKwc196yyJI/KEi00EJEzV7Zn4K4EVjvn3hO/fy2wxDl30ynPWwOsAZg8efLCvXsHP5Q5EIfrmln6xbX9Pi8IvPzZleTm5qakHV2HcdvaYiyfMZai/BzWbj9CS3ylX1sbzJ1YyvIZ44acvWifFzajvIDapjA19S3UNIaoPtrYbfh4qHPgGhtb+cZTB5LcO8NLeZ5x46XTaA0Pjzlwra0xyopyuWrh5I4FLyIikjkjNQM3IM65e4F7wRtCTdXfqSjJ48IppTy/t47KQh//umQSuaM6h4QiEeOis8s5Z0JZSov25uT4WTmnEqjsdnzaxcVJ/TuBgI/Zld45U5URWXPpuTz0QjWxWNuAsoMF+UGqSvPYd7xlQEHkQH+3r7itb2kjHI4xdVw+u480k5PDkM432MUZIiIiXWV7AHcAmNTl/sT4sYwwMx66cblqqiVRfn6QGy6anulmJHZephsgIiJnqmzfu+l54Gwzm2ZmOcDVwKOZbJDPZ5QXjVLwJiIiIimT1Rk451zEzG4CHscrI/I959zWDDdLREREJKWyOoADcM79DvhdptshIiIiki7ZPoQqIiIicsZRACciIiKSZRTAiYiIiGQZBXAiIiIiWUYBnIiIiEiWUQAnIiIikmUUwImIiIhkGQVwIiIiIllGAZyIiIhIllEAJyIiIpJlFMCJiIiIZBlzzmW6DWllZrXA3ky3YwDGAkcz3YhhQP3QSX3hUT90Ul941A+d1BeekdQPU5xz5acePOMCuGxhZhucc4sy3Y5MUz90Ul941A+d1Bce9UMn9YXnTOgHDaGKiIiIZBkFcCIiIiJZRgHc8HVvphswTKgfOqkvPOqHTuoLj/qhk/rCM+L7QXPgRERERLKMMnAiIiIiWUYBnIiIiEiWUQCXJmY2yczWmtk2M9tqZjfHj482sz+Y2Svx32Xx47PN7BkzazWzjyU4n9/MXjSz36T7Wk5HMvvBzPaY2ctmtsnMNmTiek5Hkvui1Mx+bmY7zGy7mS3LxDUNRbL6wcxmxd8L7T/1ZvbhTF3XUCT5PfHv8XNsMbMHzCw3E9c0FEnuh5vjfbA1294PMKS++Fcz2xz/bFxnZvO6nGu1mf3dzF41s09m6pqGIsn98D0zqzGzLZm6nqRwzuknDT9AJXBB/HYRsBM4B/gy8Mn48U8CX4rfHgdcCNwJfCzB+T4C/BT4TaavLVP9AOwBxmb6moZJX9wPvCd+OwcozfT1ZaIfupzTDxzGK4CZ8WtMd18AE4DdQF78/s+Ad2b6+jLQD3OBLUA+EAD+CMzI9PWluC+WA2Xx228Gno3f9gO7gLPinxEvAedk+vrS3Q/x+5cAFwBbMn1dp/OjDFyaOOcOOedeiN9uALbjfchegfflS/z3P8afU+Ocex5oO/VcZjYReCtwXxqanlTJ7Idsl6y+MLMSvA+k78afF3bO1aXlIpIgRe+JlcAu51w27LrSIcl9EQDyzCyAF8AcTHHzkyaJ/TAH74u72TkXAZ4E3pGGS0iaIfTFOufcifjx9cDE+O3FwKvOudecc2Hgwfg5skIS+wHn3FPA8TQ1PWUUwGWAmU0FFgDPAhXOuUPxhw4DFQM4xTeAjwOxVLQvXZLQDw54wsw2mtmalDQyTU6zL6YBtcD3zRtWv8/MClLV1lRKwnui3dXAA0ltXJqdTl845w4A/wlUA4eAk865J1LW2BQ6zffEFuBiMxtjZvnAW4BJKWpqyg2hL94N/D5+ewKwr8tj++PHss5p9sOIoQAuzcysEHgY+LBzrr7rY87L7fZZ18XM3gbUOOc2pq6VqXe6/RB3kXPuArz0+AfN7JLktzT1ktAXAbzhgHuccwuAJryhhKySpPcEZpYDXA78T9IbmSZJ+Jwow8tMTAOqgAIz+78pam7KnG4/OOe2A18CngAeAzYB0dS0NrUG2xdm9nq8wOUTaWtkGqgfOimASyMzC+K98X7inHskfviImVXGH68Eavo5zQrgcjPbg5cCf4OZ/ThFTU6JJPVDe5YB51wN8Au8IYKskqS+2A/sd849G7//c7yALmsk6z0R92bgBefckeS3NPWS1BdvBHY752qdc23AI3hzgrJGEj8nvuucW+icuwQ4gTd3KqsMti/M7Hy8KTZXOOeOxQ8foHv2cWL8WNZIUj+MGArg0sTMDG+O0nbn3Ne6PPQocH389vXAr/o6j3Pu/znnJjrnpuINE/3ZOZc1/7JOVj+YWYGZFbXfBlbhDZdkjSS+Jw4D+8xsVvzQSmBbkpubMsnqhy6uIUuHT5PYF9XAUjPLj59zJd6coayQzPeEmY2L/56MN//tp8ltbWoNti/i1/kIcK1zrmuw+jxwtplNi2epr46fIysksR9Gjr5WOOgneT/ARXip3c14afxNePMxxgB/Al7BWyE1Ov788XiZlXqgLn67+JRzXkr2rUJNSj/graR6Kf6zFfh0pq8tk+8JYD6wIX6uXxJffZUNP0nuhwLgGFCS6esaBn1xO7AD7x82PwJGZfr6MtQPT+P9g+YlYGWmry0NfXEfXqax/bkbupzrLXgZyF3Z9pmZ5H54AG9uaFv8vfLuTF/fUH60lZaIiIhIltEQqoiIiEiWUQAnIiIikmUUwImIiIhkGQVwIiIiIllGAZyIiIhIllEAJyLSCzOLmtkmM9tqZi+Z2UfNrM/PTTObamb/kq42isiZSQGciEjvWpxz851z5wKX4e3ycFs/r5kKKIATkZRSHTgRkV6YWaNzrrDL/bPwKtqPBabgFcgtiD98k3NunZmtB+YAu4H7gf8CvohXeHsU8C3n3HfSdhEiMiIpgBMR6cWpAVz8WB0wC2gAYs65kJmdDTzgnFtkZpcCH3POvS3+/DXAOOfcHWY2CvgbcJVzbndaL0ZERpRAphsgIpKlgsBdZjYfiAIze3neKuB8M7syfr8EOBsvQyciMiQK4EREBig+hBoFavDmwh0B5uHNJw719jLgQ865x9PSSBE5I2gRg4jIAJhZOfBt4C7nzT0pAQ4552LAtYA//tQGoKjLSx8H3m9mwfh5ZppZASIip0EZOBGR3uWZ2Sa84dII3qKFr8Ufuxt42MyuAx4DmuLHNwNRM3sJ+AHwTbyVqS+YmQG1wD+m6wJEZGTSIgYRERGRLKMhVBEREZEsowBOREREJMsogBMRERHJMgrgRERERLKMAjgRERGRLKMATkRERCTLKIATERERyTL/CxiR/HPM2fguAAAAAElFTkSuQmCC\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Creating a plotting function for time series data"
      ],
      "metadata": {
        "id": "nCdicnOKR-sF"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Creating a function to plot time series data\n",
        "\n",
        "def plot_time_series(timesteps, values, format='.', start=0, end=None, label=None):\n",
        "  \"\"\"\n",
        "  Plots a timesteps (a series of points in time) against values (a series of values across timesteps).\n",
        "  \n",
        "  Parameters\n",
        "  ---------\n",
        "  timesteps : array of timesteps\n",
        "  values : array of values across time\n",
        "  format : style of plot, default \".\"\n",
        "  start : where to start the plot (setting a value will index from start of timesteps & values)\n",
        "  end : where to end the plot (setting a value will index from end of timesteps & values)\n",
        "  label : label to show on plot of values\n",
        "  \"\"\"\n",
        "  # Plot the series\n",
        "  plt.plot(timesteps[start:end], values[start:end], format, label=label)\n",
        "  plt.xlabel(\"Time\")\n",
        "  plt.ylabel(\"BTC Price\")\n",
        "  if label:\n",
        "    plt.legend(fontsize=14) # make label bigger\n",
        "  plt.grid(True)"
      ],
      "metadata": {
        "id": "4cE5im9_RjW6"
      },
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Try out our plotting function\n",
        "plt.figure(figsize=(10, 7))\n",
        "plot_time_series(timesteps=X_train, values=y_train, label=\"Train data\")\n",
        "plot_time_series(timesteps=X_test, values=y_test, label=\"Test data\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 442
        },
        "id": "iP5uCpanT1vU",
        "outputId": "0a40e614-555d-4ac4-bfd0-23e34c646514"
      },
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x504 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eotxQI8XI5h5",
        "outputId": "a38be3b5-6c0f-41fd-bb17-61c786244f27"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Modelling Experiments"
      ],
      "metadata": {
        "id": "JURD8u4cZHh7"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "\n",
        "Bitcoin historical data is given, to model it, I'll run a series of modelling experiments and see which model performs best.\n",
        "\n",
        "Terms to be familiar with:\n",
        "\n",
        "* **Horizon** = number of timesteps into the future that are going to be predicted\n",
        "* **Window size** = number of timesteps that are going to be used to predict horizon\n",
        "Modelling experiments:\n",
        "\n",
        "* 0 Naïve model (baseline)\n",
        "* 1 Dense model, horizon = 1, window = 7\n",
        "* 2 Same as 1, horizon = 1, window = 30\n",
        "* 3 Same as 1, horizon = 7, window = 30\n",
        "* 4 LSTM\n",
        "* 5 Ensemble (multiple models optimized on different loss functions)\n",
        "* 6 Same as 1 (but with turkey data introduced)"
      ],
      "metadata": {
        "id": "169s7m0XZEf_"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Model 0: Naive Forecast Model"
      ],
      "metadata": {
        "id": "uF5mFG6RuK0Z"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "The prediction at timestep t (y-hat) is equal to the value at timestep t-1 (previous timestep) : (this is for a horizon=1)"
      ],
      "metadata": {
        "id": "Z_Xf-cXp5MSs"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "  #Creating a Naive Forecast\n",
        "\n",
        "  naive_forecast = y_test[:-1] #offsetting y_test by 1\n",
        "\n",
        "  naive_forecast[:10], naive_forecast[-10:]"
      ],
      "metadata": {
        "id": "zCyaiYf6U4YQ",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "aea4ef95-2fe8-4130-da59-63110961f84d"
      },
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([9226.485821, 8794.358645, 8798.042055, 9081.186878, 8711.534339,\n",
              "        8760.892718, 8749.520591, 8656.970922, 8500.643558, 8469.260899]),\n",
              " array([57107.12067, 58788.20968, 58102.19143, 55715.54665, 56573.55547,\n",
              "        52147.82119, 49764.13208, 50032.69314, 47885.62525, 45604.61575]))"
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Plot naive forecast\n",
        "plt.figure(figsize=(10,7))\n",
        "# plot_time_series(timesteps = X_train, values = y_train, label=\"Train Data\")\n",
        "plot_time_series(timesteps = X_test, values = y_test, start=350, format=\"-\", label=\"Test Data\")\n",
        "plot_time_series(timesteps = X_test[1:], values = naive_forecast, start=350, format =\"-\", label=\"Naive Forecast\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 442
        },
        "id": "r1aC6Plp6Jkt",
        "outputId": "89015fc0-9f1a-4c1c-f43b-32759d188cfe"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x504 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Evaluating a time series model\n",
        "\n",
        "In the end a number has to be predicted, that means it is a regression problem, so, some regression-like metrics are needed.\n",
        "\n",
        "A few common regression metrics (which can also be used for time series forecasting):\n",
        "* MAE - mean absolute error\n",
        "* MSE - mean squared error\n",
        "* RMSE - root mean square error\n",
        "* MAPE/sMAPE - (symmetric) mean absolute percentage error\n",
        "* MASE - mean absolute scaled error\n",
        "\n",
        "For all of the above metrics, **lower is better**, for example, an MAE of 0 that is better than an MAE of 100.\n",
        "\n",
        "The main thing to evaluate here is: **how do our model's forecasts (y_pred) compare against the actual values (y_true or ground truth values)?**\n"
      ],
      "metadata": {
        "id": "hdAjGL_79EKy"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import tensorflow as tf"
      ],
      "metadata": {
        "id": "u4yqG82A7can"
      },
      "execution_count": 21,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# MASE Implementation since it is not in the tf.keras.metrics class\n",
        "# Seasonality is a characteristic of a time series in which the data experiences regular and predictable changes that recur every calendar year.\n",
        "\n",
        "def mean_absolute_scaled_error(y_true, y_pred):\n",
        "  \"\"\"\n",
        "  Implement MASE (assuming no seasonality of data).\n",
        "  \"\"\"\n",
        "  mae = tf.reduce_mean(tf.abs(y_true - y_pred))\n",
        "\n",
        "  # Find MAE of naive forecast (no seasonality)\n",
        "  mae_naive_no_season = tf.reduce_mean(tf.abs(y_true[1:] - y_true[:-1])) # our seasonality is 1 day (hence the shifting of 1 day)\n",
        "\n",
        "  return mae / mae_naive_no_season"
      ],
      "metadata": {
        "id": "wfc_yJs89DIF"
      },
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "mean_absolute_scaled_error(y_true = y_test[1:], y_pred=naive_forecast).numpy()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "A1u28Hs1lpRS",
        "outputId": "8e7fcb9a-3fd9-4f06-f9fc-8fa537ef127f"
      },
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.9995699939172406"
            ]
          },
          "metadata": {},
          "execution_count": 23
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Above it can be seen that MASE for Naive Forecast is almost 1, which should be the case for Naive Forecast"
      ],
      "metadata": {
        "id": "q28elSBpmdVj"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "In practice, if the function is correct, the naïve model should achieve an MASE of 1 (or very close to 1). Any model worse than the naïve forecast will achieve an MASE of >1 and any model better than the naïve forecast will achieve an MASE of <1."
      ],
      "metadata": {
        "id": "_7ynX9hWrnnh"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Putting each different evaluation metrics together into a function."
      ],
      "metadata": {
        "id": "YGvr-rHlrsdZ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "y_test[0].dtype"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8QlNs9gCnZGJ",
        "outputId": "5656d7ea-209a-471e-e77a-b717560aef34"
      },
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "dtype('float64')"
            ]
          },
          "metadata": {},
          "execution_count": 24
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def evaluate_preds(y_true, y_pred):\n",
        "  # Make sure float32 (for metric calculations), if not in float32 tensorflow sometimes gives errors\n",
        "  y_true = tf.cast(y_true, dtype=tf.float32)\n",
        "  y_pred = tf.cast(y_pred, dtype=tf.float32)\n",
        "\n",
        "  # Calculate various metrics\n",
        "  mae = tf.keras.metrics.mean_absolute_error(y_true, y_pred)\n",
        "  mse = tf.keras.metrics.mean_squared_error(y_true, y_pred) # puts and emphasis on outliers (all errors get squared)\n",
        "  rmse = tf.sqrt(mse)\n",
        "  mape = tf.keras.metrics.mean_absolute_percentage_error(y_true, y_pred)\n",
        "  mase = mean_absolute_scaled_error(y_true, y_pred)\n",
        "  \n",
        "  return {\"mae\": mae.numpy(),\n",
        "          \"mse\": mse.numpy(),\n",
        "          \"rmse\": rmse.numpy(),\n",
        "          \"mape\": mape.numpy(),\n",
        "          \"mase\": mase.numpy()}"
      ],
      "metadata": {
        "id": "2sgVJrot_17v"
      },
      "execution_count": 25,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Testing function \n",
        "\n",
        "naive_results = evaluate_preds(y_true=y_test[1:], y_pred=naive_forecast)\n",
        "\n",
        "naive_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KbKM4CgZusWy",
        "outputId": "c4b61f1f-5b9d-4162-f51f-3838985dbce9"
      },
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'mae': 567.9802,\n",
              " 'mape': 2.516525,\n",
              " 'mase': 0.99957,\n",
              " 'mse': 1147547.0,\n",
              " 'rmse': 1071.2362}"
            ]
          },
          "metadata": {},
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Find average price of bitcoin\n",
        "\n",
        "tf.reduce_mean(y_test).numpy()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OIzF7N0VwIlZ",
        "outputId": "ff9f4ce7-c182-4203-cdd6-6a2e3bf0c522"
      },
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "20056.632963623655"
            ]
          },
          "metadata": {},
          "execution_count": 27
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "- Here we can see that the average price of bitcoin in test dateset is 20,056 dollars\n",
        "- So, we can say that on an average each prediction in naive forecast is off by 567.98 dollars"
      ],
      "metadata": {
        "id": "fcXXBVDBx5tZ"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Format Data Part 2: Windowing Dataset\n",
        "\n",
        "\n",
        "Windowing is a method to turn a time series dataset into supervised learning problem.\n",
        "\n",
        "\n",
        "Window for one week with the target of predicting the next day (Bitcoin prices)\n",
        "\n",
        "- [123.654, 125.455, 108.584, 118.674, 121.338, 120.655, 121.795] -> [123.033]\n",
        "- [125.455, 108.584, 118.674, 121.338, 120.655, 121.795, 123.033] -> [124.049]\n",
        "- [108.584, 118.674, 121.338, 120.655, 121.795, 123.033, 124.049] -> [125.961]"
      ],
      "metadata": {
        "id": "nF19FaSB0xE2"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Setting up global variables\n",
        "HORIZON = 1 # predict one step at a time\n",
        "WINDOW_SIZE = 7 # use one week worth of timesteps to predict the horizon"
      ],
      "metadata": {
        "id": "HgwlFrovx4wf"
      },
      "execution_count": 28,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "Writing a function to take in an array and turn it into a window and horizon."
      ],
      "metadata": {
        "id": "DMZcCts22sAz"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Create a function to label windowed data  \n",
        "def get_labelled_windows(x, horizon=1):\n",
        "\n",
        "  \"\"\"\n",
        "  Creates labels for windowed dataset.\n",
        "\n",
        "  E.g. if horizon=1 (default)\n",
        "  Input: [1, 2, 3, 4, 5, 6] -> Output: ([1, 2, 3, 4, 5], [6])\n",
        "  \"\"\"\n",
        "  return x[:, :-horizon], x[:, -horizon:]\n",
        "  "
      ],
      "metadata": {
        "id": "z0p69DL32rY7"
      },
      "execution_count": 29,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Testing out the window labelling function\n",
        "\n",
        "test_window, test_label = get_labelled_windows(tf.expand_dims(tf.range(8)+1, axis=0))\n",
        "test_window, test_label\n",
        "print(f\"Window: {tf.squeeze(test_window).numpy()} -> Label:  {tf.squeeze(test_label).numpy()} \")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "encnnV3ksjAZ",
        "outputId": "23c90fe8-644a-4d6e-deaa-d8f38da1d743"
      },
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Window: [1 2 3 4 5 6 7] -> Label:  8 \n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Function to make windows for an entire time series:\n",
        "\n",
        "1. Creates a window step of specific window size, for example: [[0, 1, 2, 3, 4, 5, 6, 7]]\n",
        "2. Uses NumPy indexing to create a 2D of multiple window steps, for example:\n",
        "```\n",
        "  - [[0, 1, 2, 3, 4, 5, 6, 7],\n",
        "  - [1, 2, 3, 4, 5, 6, 7, 8],\n",
        "  - [2, 3, 4, 5, 6, 7, 8, 9]]\n",
        "```\n",
        "3. Uses the 2D array of multiple window steps (from 2.) to index on a target series (e.g. historical price of bitcoin)\n",
        "4. Use the get_labelled_windows() function created above to turn the window steps into windows with a specified horizon"
      ],
      "metadata": {
        "id": "HSM-emnm3jSF"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Create a function to view numpy arrays as windows\n",
        "\n",
        "def make_windows(x, window_size=WINDOW_SIZE, horizon = HORIZON):\n",
        "  \"\"\"\n",
        "  Turns a 1D array into a 2D array of sequential lebelled windows of window_size\n",
        "  \"\"\"\n",
        "  # 1. Create a window of a specific window_size (add the horizon on the end for labelling later)\n",
        "  window_step = np.expand_dims(np.arange(window_size+horizon), axis = 0)\n",
        "  print(f\"Window step:\\n  {window_step}\")\n",
        "\n",
        "  # 2. Create a 2D array of windows of window_size (minus 1 to account for 0 indexing) using numpy indexing\n",
        "  window_indexes = window_step + np.expand_dims(np.arange(len(x) - (window_size+horizon-1)), axis=0).T\n",
        "  print(f\"Window indexes:\\n {window_indexes[:3], window_indexes[-3:], window_indexes.shape}\")\n",
        "\n",
        "  # 3. Index on target array (time series) with 2D array of multiple window steps\n",
        "  windowed_array = x[window_indexes]\n",
        "  print(f\"Windowed array:\\n {windowed_array}\")\n",
        "\n",
        "  # 4. Get the labelled windows\n",
        "  windows, labels = get_labelled_windows(windowed_array, horizon=horizon)\n",
        "\n",
        "  return windows, labels\n",
        "  \n"
      ],
      "metadata": {
        "id": "SEouwtaQ3J-W"
      },
      "execution_count": 31,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "full_windows, full_labels = make_windows(prices, window_size = WINDOW_SIZE, horizon = HORIZON)\n",
        "print(f\"Lenght of full windows: {len(full_windows)} \\nLength of full labels: {len(full_labels)}\")\n",
        "len(prices)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VrtyKH1Q9jwc",
        "outputId": "172a3a18-3b0c-499e-9cdb-05652ce5492b"
      },
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Window step:\n",
            "  [[0 1 2 3 4 5 6 7]]\n",
            "Window indexes:\n",
            " (array([[0, 1, 2, 3, 4, 5, 6, 7],\n",
            "       [1, 2, 3, 4, 5, 6, 7, 8],\n",
            "       [2, 3, 4, 5, 6, 7, 8, 9]]), array([[2777, 2778, 2779, 2780, 2781, 2782, 2783, 2784],\n",
            "       [2778, 2779, 2780, 2781, 2782, 2783, 2784, 2785],\n",
            "       [2779, 2780, 2781, 2782, 2783, 2784, 2785, 2786]]), (2780, 8))\n",
            "Windowed array:\n",
            " [[  123.65499   125.455     108.58483 ...   120.65533   121.795\n",
            "    123.033  ]\n",
            " [  125.455     108.58483   118.67466 ...   121.795     123.033\n",
            "    124.049  ]\n",
            " [  108.58483   118.67466   121.33866 ...   123.033     124.049\n",
            "    125.96116]\n",
            " ...\n",
            " [58788.20968 58102.19143 55715.54665 ... 49764.13208 50032.69314\n",
            "  47885.62525]\n",
            " [58102.19143 55715.54665 56573.55547 ... 50032.69314 47885.62525\n",
            "  45604.61575]\n",
            " [55715.54665 56573.55547 52147.82119 ... 47885.62525 45604.61575\n",
            "  43144.47129]]\n",
            "Lenght of full windows: 2780 \n",
            "Length of full labels: 2780\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2787"
            ]
          },
          "metadata": {},
          "execution_count": 32
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# View the first 3 windows/labels\n",
        "\n",
        "for i in range(3):\n",
        "  print(f\"Window: {full_windows[i]} -> Label: {full_labels[i]}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vqPYuGnJzSxH",
        "outputId": "24bba701-fafc-4d1e-81e7-3c4dc18cf49d"
      },
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Window: [123.65499 125.455   108.58483 118.67466 121.33866 120.65533 121.795  ] -> Label: [123.033]\n",
            "Window: [125.455   108.58483 118.67466 121.33866 120.65533 121.795   123.033  ] -> Label: [124.049]\n",
            "Window: [108.58483 118.67466 121.33866 120.65533 121.795   123.033   124.049  ] -> Label: [125.96116]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "prices[:15]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TWJOzmSY3J0c",
        "outputId": "219ca1cf-bade-4e6e-a11d-7d315a60963b"
      },
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([123.65499, 125.455  , 108.58483, 118.67466, 121.33866, 120.65533,\n",
              "       121.795  , 123.033  , 124.049  , 125.96116, 125.27966, 125.9275 ,\n",
              "       126.38333, 135.24199, 133.20333])"
            ]
          },
          "metadata": {},
          "execution_count": 34
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# View the last 3 windows/labels\n",
        "\n",
        "for i in range(3):\n",
        "  print(f\"Window: {full_windows[i-3]} -> Label: {full_labels[i-3]}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7NKAUc0B3Vya",
        "outputId": "200ccb76-d2d8-4f40-d1c5-ab6395064e1e"
      },
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Window: [58788.20968 58102.19143 55715.54665 56573.55547 52147.82119 49764.13208\n",
            " 50032.69314] -> Label: [47885.62525]\n",
            "Window: [58102.19143 55715.54665 56573.55547 52147.82119 49764.13208 50032.69314\n",
            " 47885.62525] -> Label: [45604.61575]\n",
            "Window: [55715.54665 56573.55547 52147.82119 49764.13208 50032.69314 47885.62525\n",
            " 45604.61575] -> Label: [43144.47129]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "prices[-15:]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JX7DNmDB3436",
        "outputId": "7b758043-053c-49ed-c879-9c636577eb8f"
      },
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([57255.30684, 53658.84312, 57252.70218, 56583.84988, 57107.12067,\n",
              "       58788.20968, 58102.19143, 55715.54665, 56573.55547, 52147.82119,\n",
              "       49764.13208, 50032.69314, 47885.62525, 45604.61575, 43144.47129])"
            ]
          },
          "metadata": {},
          "execution_count": 36
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Turning windows into training and test tests\n",
        "\n"
      ],
      "metadata": {
        "id": "dtkM0K1k4JMj"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Make the test/train splits\n",
        "def make_train_test_splits(windows, labels, test_split=0.2):\n",
        "  \"\"\"\n",
        "  Splits, matching pairs of windows and labels into train and test splits\n",
        "  \"\"\"\n",
        "  split_size = int(len(windows) * (1-test_split)) # this will default to 80% train and 20% test\n",
        "  #print(split_size)\n",
        "  train_windows = windows[:split_size]\n",
        "  train_labels = labels[:split_size]\n",
        "  test_windows = windows[split_size:]\n",
        "  test_labels = labels[split_size:]\n",
        "\n",
        "  return train_windows, test_windows, train_labels, test_labels"
      ],
      "metadata": {
        "id": "OCv3e1I64AkR"
      },
      "execution_count": 37,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Create train and test windows\n",
        "train_windows, test_windows, train_labels, test_labels = make_train_test_splits(full_windows, full_labels)\n",
        "len(train_windows), len(test_windows), len(train_labels), len(test_labels)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bS0sZ0je6CUZ",
        "outputId": "1869b4d0-067a-4e0c-a580-606e3b598441"
      },
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(2224, 556, 2224, 556)"
            ]
          },
          "metadata": {},
          "execution_count": 38
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "len(full_windows) * 0.8"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9se3s7dv65Vv",
        "outputId": "574a5771-9b21-4c1a-f39a-17083b131210"
      },
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2224.0"
            ]
          },
          "metadata": {},
          "execution_count": 39
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train_windows[:5], train_labels[:5]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-x6PZxDv7MJR",
        "outputId": "8593e157-0979-4428-a4c3-762a25098ca7"
      },
      "execution_count": 40,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([[123.65499, 125.455  , 108.58483, 118.67466, 121.33866, 120.65533,\n",
              "         121.795  ],\n",
              "        [125.455  , 108.58483, 118.67466, 121.33866, 120.65533, 121.795  ,\n",
              "         123.033  ],\n",
              "        [108.58483, 118.67466, 121.33866, 120.65533, 121.795  , 123.033  ,\n",
              "         124.049  ],\n",
              "        [118.67466, 121.33866, 120.65533, 121.795  , 123.033  , 124.049  ,\n",
              "         125.96116],\n",
              "        [121.33866, 120.65533, 121.795  , 123.033  , 124.049  , 125.96116,\n",
              "         125.27966]]), array([[123.033  ],\n",
              "        [124.049  ],\n",
              "        [125.96116],\n",
              "        [125.27966],\n",
              "        [125.9275 ]]))"
            ]
          },
          "metadata": {},
          "execution_count": 40
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "test_windows[:5], test_labels[:5]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bm2LlcSC7VoJ",
        "outputId": "3d3b1811-f980-4108-bb97-08d0b5197728"
      },
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([[9290.896602, 9202.415451, 9369.628081, 9326.599624, 9335.752402,\n",
              "         9226.485821, 8794.358645],\n",
              "        [9202.415451, 9369.628081, 9326.599624, 9335.752402, 9226.485821,\n",
              "         8794.358645, 8798.042055],\n",
              "        [9369.628081, 9326.599624, 9335.752402, 9226.485821, 8794.358645,\n",
              "         8798.042055, 9081.186878],\n",
              "        [9326.599624, 9335.752402, 9226.485821, 8794.358645, 8798.042055,\n",
              "         9081.186878, 8711.534339],\n",
              "        [9335.752402, 9226.485821, 8794.358645, 8798.042055, 9081.186878,\n",
              "         8711.534339, 8760.892718]]), array([[8798.042055],\n",
              "        [9081.186878],\n",
              "        [8711.534339],\n",
              "        [8760.892718],\n",
              "        [8749.520591]]))"
            ]
          },
          "metadata": {},
          "execution_count": 41
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Check to see if train labels are the same (before and after the window split)\n",
        "\n",
        "np.array_equal(np.squeeze(train_labels[:-HORIZON-1]), y_train[WINDOW_SIZE:])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mxbKb8457bKj",
        "outputId": "93563752-8860-43f4-8f05-caf89aa7e1f7"
      },
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "metadata": {},
          "execution_count": 42
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Make a Modelling Checkpoint\n",
        "\n",
        "Because the model's performance will fluctuate from experiment to experiment, the comparison should be done in a way in which each model's best performance is compared against each model's best performance.\n",
        "\n",
        "For example, if ```model_1``` performed incredibly well on epoch 55 but its performance fell off toward epoch 100, I want the version of the model from epoch 55 to compare to other models rather than the version of the model from epoch 100.\n",
        "\n",
        "For this: implementing a ```ModelCheckpoint``` callback."
      ],
      "metadata": {
        "id": "Ey9GOIT18eyw"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import os\n",
        "\n",
        "# Create a function to implement a ModelCheckPoint callback with a specific filename\n",
        "def create_model_checkpoint(model_name, save_path = \"/content/drive/MyDrive/BTC Predictor/model_experiments\"):\n",
        "   return tf.keras.callbacks.ModelCheckpoint(filepath=os.path.join(save_path, model_name), \n",
        "                                             monitor = 'val_loss', \n",
        "                                             verbose=0, # only output a limited amount of text,\n",
        "                                             save_best_only = True) "
      ],
      "metadata": {
        "id": "e5YMs1SU8PYD"
      },
      "execution_count": 43,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Model 1: Dense Model (window =7, horizon =1)\n",
        "\n",
        "* A single dense layer with 128 units and ReLu\n",
        "* An output layer with linear activation (no activation) - (here don't use softmax function as we only to predict a number)\n",
        "* Adam optimization and MAE Loss function \n",
        "* And as dataset is small we'll use a batch size of 128 and 100 epochs \n",
        "\n",
        "***\n",
        "* We add a Dense layer of 128 neurons (each neuron is connected to each neuron in previous and next dense layers)\n",
        "* In a neural network, the activation function is responsible for transforming the summed weighted input from the node into the activation of the node or output for that input.\n",
        "* The rectified linear activation (ReLu) function or ReLU is a piecewise linear function that will output the input directly if it is positive, otherwise, it will output zero."
      ],
      "metadata": {
        "id": "dOvvz3WHneib"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import tensorflow as tf\n",
        "from tensorflow.keras import layers\n",
        "\n",
        "# Set random seed for as reproducible results as possible\n",
        "tf.random.set_seed(42)\n",
        "\n",
        "# 1. Construct model\n",
        "model_1 = tf.keras.Sequential([\n",
        "    layers.Dense(128, activation = \"relu\"), \n",
        "    layers.Dense(HORIZON, activation = \"linear\")],   # here HORIZON = 1 which is our output neuron, linear activation is the same as having no activation\n",
        "    name = \"model_1_dense\")                          # name our model so we can save it\n",
        "\n",
        "# 2. Compile model\n",
        "model_1.compile(loss = \"mae\", \n",
        "                optimizer = tf.keras.optimizers.Adam(),\n",
        "                metrics = [\"mae\", \"mse\"]\n",
        "                 )\n",
        "# 3. Fit model\n",
        "model_1.fit(x= train_windows,\n",
        "            y= train_labels,\n",
        "            epochs = 100,\n",
        "            verbose = 1, # as we don't want to print much data\n",
        "            batch_size=128  , \n",
        "            validation_data = [test_windows, test_labels], # after fitting training data\n",
        "            callbacks = [create_model_checkpoint(model_1.name)]\n",
        "            )\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KktSl-5r-Otc",
        "outputId": "68438c6a-2b55-494b-c9d0-cb26912ddfce"
      },
      "execution_count": 44,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            " 1/18 [>.............................] - ETA: 46s - loss: 2111.0571 - mae: 2111.0571 - mse: 9255782.0000INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 3s 45ms/step - loss: 780.3455 - mae: 780.3455 - mse: 2312725.7500 - val_loss: 2279.6528 - val_mae: 2279.6528 - val_mse: 12772731.0000\n",
            "Epoch 2/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 348.8859 - mae: 348.8859 - mse: 600802.6250INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 1s 29ms/step - loss: 247.6756 - mae: 247.6756 - mse: 285481.1875 - val_loss: 1005.9993 - val_mae: 1005.9993 - val_mse: 3441833.2500\n",
            "Epoch 3/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 134.4855 - mae: 134.4855 - mse: 70765.6953INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 0s 28ms/step - loss: 188.4116 - mae: 188.4116 - mse: 171891.6094 - val_loss: 923.2862 - val_mae: 923.2861 - val_mse: 2868220.7500\n",
            "Epoch 4/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 211.3724 - mae: 211.3724 - mse: 238018.8906INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 2s 108ms/step - loss: 169.4340 - mae: 169.4340 - mse: 151700.5781 - val_loss: 900.5872 - val_mae: 900.5872 - val_mse: 2683715.5000\n",
            "Epoch 5/100\n",
            "15/18 [========================>.....] - ETA: 0s - loss: 168.5602 - mae: 168.5602 - mse: 153500.3281INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 1s 40ms/step - loss: 165.0894 - mae: 165.0894 - mse: 145490.7812 - val_loss: 895.2237 - val_mae: 895.2237 - val_mse: 2564574.5000\n",
            "Epoch 6/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 117.3681 - mae: 117.3681 - mse: 55147.0820INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 0s 29ms/step - loss: 158.5210 - mae: 158.5210 - mse: 133817.7344 - val_loss: 855.1984 - val_mae: 855.1984 - val_mse: 2481873.2500\n",
            "Epoch 7/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 180.3349 - mae: 180.3349 - mse: 241586.4219INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 1s 30ms/step - loss: 151.3566 - mae: 151.3566 - mse: 123446.9766 - val_loss: 840.9166 - val_mae: 840.9166 - val_mse: 2276451.5000\n",
            "Epoch 8/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 121.1138 - mae: 121.1138 - mse: 57999.1211INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 1s 37ms/step - loss: 145.2560 - mae: 145.2560 - mse: 117694.5000 - val_loss: 803.5956 - val_mae: 803.5956 - val_mse: 2140573.5000\n",
            "Epoch 9/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 126.2215 - mae: 126.2215 - mse: 75301.4141INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 0s 28ms/step - loss: 144.3546 - mae: 144.3546 - mse: 115269.3047 - val_loss: 799.5454 - val_mae: 799.5454 - val_mse: 2049207.7500\n",
            "Epoch 10/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 148.4502 - mae: 148.4502 - mse: 164283.9375INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 0s 29ms/step - loss: 141.2943 - mae: 141.2943 - mse: 112394.5938 - val_loss: 763.5010 - val_mae: 763.5010 - val_mse: 1933140.7500\n",
            "Epoch 11/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 135.6595 - mae: 135.6595 - mse: 104118.2344 - val_loss: 771.3356 - val_mae: 771.3356 - val_mse: 1885849.7500\n",
            "Epoch 12/100\n",
            "18/18 [==============================] - 0s 5ms/step - loss: 134.1700 - mae: 134.1700 - mse: 101780.8672 - val_loss: 782.8079 - val_mae: 782.8079 - val_mse: 1881727.1250\n",
            "Epoch 13/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 134.6015 - mae: 134.6015 - mse: 102345.2266 - val_loss: 784.4449 - val_mae: 784.4449 - val_mse: 1860996.1250\n",
            "Epoch 14/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 136.5593 - mae: 136.5593 - mse: 78599.7188INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 0s 28ms/step - loss: 130.6127 - mae: 130.6127 - mse: 96225.1328 - val_loss: 751.3234 - val_mae: 751.3233 - val_mse: 1749609.0000\n",
            "Epoch 15/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 131.5350 - mae: 131.5350 - mse: 65715.5938INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 1s 45ms/step - loss: 128.8347 - mae: 128.8347 - mse: 94306.9922 - val_loss: 696.5756 - val_mae: 696.5756 - val_mse: 1606352.3750\n",
            "Epoch 16/100\n",
            "18/18 [==============================] - 0s 12ms/step - loss: 124.7739 - mae: 124.7739 - mse: 90394.9375 - val_loss: 702.4698 - val_mae: 702.4698 - val_mse: 1592416.2500\n",
            "Epoch 17/100\n",
            "18/18 [==============================] - 0s 5ms/step - loss: 123.4474 - mae: 123.4474 - mse: 89247.7031 - val_loss: 704.9239 - val_mae: 704.9239 - val_mse: 1580782.7500\n",
            "Epoch 18/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 137.2540 - mae: 137.2540 - mse: 95732.8359INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 1s 34ms/step - loss: 122.2105 - mae: 122.2105 - mse: 87607.5938 - val_loss: 667.9724 - val_mae: 667.9724 - val_mse: 1498881.6250\n",
            "Epoch 19/100\n",
            "18/18 [==============================] - 0s 14ms/step - loss: 121.7263 - mae: 121.7263 - mse: 86632.2422 - val_loss: 718.8796 - val_mae: 718.8796 - val_mse: 1589299.2500\n",
            "Epoch 20/100\n",
            "10/18 [===============>..............] - ETA: 0s - loss: 125.0261 - mae: 125.0261 - mse: 93809.5234 INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 1s 39ms/step - loss: 119.2420 - mae: 119.2420 - mse: 84334.3203 - val_loss: 657.0666 - val_mae: 657.0666 - val_mse: 1447501.1250\n",
            "Epoch 21/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 129.2764 - mae: 129.2764 - mse: 78635.7812INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 1s 54ms/step - loss: 121.2275 - mae: 121.2275 - mse: 85743.3828 - val_loss: 637.0331 - val_mae: 637.0331 - val_mse: 1401774.1250\n",
            "Epoch 22/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 119.9544 - mae: 119.9544 - mse: 84050.1094 - val_loss: 671.2487 - val_mae: 671.2487 - val_mse: 1447895.6250\n",
            "Epoch 23/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 116.0426 - mae: 116.0426 - mse: 57999.4492INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 0s 27ms/step - loss: 121.9248 - mae: 121.9248 - mse: 83746.1250 - val_loss: 633.3591 - val_mae: 633.3591 - val_mse: 1367377.8750\n",
            "Epoch 24/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 124.4802 - mae: 124.4802 - mse: 83062.6094INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 1s 29ms/step - loss: 116.3665 - mae: 116.3665 - mse: 81446.6562 - val_loss: 624.4852 - val_mae: 624.4852 - val_mse: 1348960.7500\n",
            "Epoch 25/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 123.8925 - mae: 123.8925 - mse: 61388.7969INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 0s 27ms/step - loss: 114.6816 - mae: 114.6816 - mse: 78912.8359 - val_loss: 619.7570 - val_mae: 619.7570 - val_mse: 1351089.7500\n",
            "Epoch 26/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 110.2667 - mae: 110.2667 - mse: 60073.7930INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 1s 29ms/step - loss: 116.4455 - mae: 116.4455 - mse: 79365.5938 - val_loss: 615.6364 - val_mae: 615.6364 - val_mse: 1321514.3750\n",
            "Epoch 27/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 116.5868 - mae: 116.5868 - mse: 78672.9062 - val_loss: 615.9629 - val_mae: 615.9629 - val_mse: 1313077.6250\n",
            "Epoch 28/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 102.5480 - mae: 102.5480 - mse: 64013.4844INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 1s 39ms/step - loss: 113.4691 - mae: 113.4691 - mse: 77440.1016 - val_loss: 608.0921 - val_mae: 608.0921 - val_mse: 1309775.3750\n",
            "Epoch 29/100\n",
            "18/18 [==============================] - 0s 5ms/step - loss: 113.7598 - mae: 113.7598 - mse: 78506.9531 - val_loss: 621.9305 - val_mae: 621.9305 - val_mse: 1313324.6250\n",
            "Epoch 30/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 124.9585 - mae: 124.9585 - mse: 84341.5859INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 0s 29ms/step - loss: 116.8613 - mae: 116.8613 - mse: 79236.4609 - val_loss: 604.4056 - val_mae: 604.4056 - val_mse: 1283407.8750\n",
            "Epoch 31/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 111.9375 - mae: 111.9375 - mse: 74849.0703 - val_loss: 609.3882 - val_mae: 609.3882 - val_mse: 1284506.2500\n",
            "Epoch 32/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 107.0553 - mae: 107.0553 - mse: 70567.9609INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 1s 30ms/step - loss: 112.4175 - mae: 112.4175 - mse: 77178.5859 - val_loss: 603.0588 - val_mae: 603.0588 - val_mse: 1273059.5000\n",
            "Epoch 33/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 112.6697 - mae: 112.6697 - mse: 75787.9766 - val_loss: 645.6973 - val_mae: 645.6973 - val_mse: 1353255.7500\n",
            "Epoch 34/100\n",
            "18/18 [==============================] - 0s 5ms/step - loss: 111.9867 - mae: 111.9867 - mse: 77245.7266 - val_loss: 604.7634 - val_mae: 604.7634 - val_mse: 1289360.3750\n",
            "Epoch 35/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 103.6643 - mae: 103.6643 - mse: 56614.1367INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 1s 29ms/step - loss: 110.9451 - mae: 110.9451 - mse: 75301.7812 - val_loss: 593.4648 - val_mae: 593.4648 - val_mse: 1250491.3750\n",
            "Epoch 36/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 114.4816 - mae: 114.4816 - mse: 76138.5391 - val_loss: 608.0074 - val_mae: 608.0074 - val_mse: 1269272.3750\n",
            "Epoch 37/100\n",
            "18/18 [==============================] - 0s 5ms/step - loss: 110.2016 - mae: 110.2016 - mse: 74635.3047 - val_loss: 597.2308 - val_mae: 597.2308 - val_mse: 1248474.0000\n",
            "Epoch 38/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 112.2372 - mae: 112.2372 - mse: 75821.6953 - val_loss: 637.9798 - val_mae: 637.9798 - val_mse: 1366130.8750\n",
            "Epoch 39/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 131.5377 - mae: 131.5377 - mse: 121641.0000INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 1s 30ms/step - loss: 115.1289 - mae: 115.1289 - mse: 76339.6797 - val_loss: 587.4680 - val_mae: 587.4680 - val_mse: 1233512.3750\n",
            "Epoch 40/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 110.0854 - mae: 110.0854 - mse: 73528.4766 - val_loss: 592.7117 - val_mae: 592.7117 - val_mse: 1251474.6250\n",
            "Epoch 41/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 110.6344 - mae: 110.6344 - mse: 73528.7891 - val_loss: 593.8997 - val_mae: 593.8997 - val_mse: 1253678.5000\n",
            "Epoch 42/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 113.5762 - mae: 113.5762 - mse: 77076.7031 - val_loss: 636.3674 - val_mae: 636.3674 - val_mse: 1317283.5000\n",
            "Epoch 43/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 116.2285 - mae: 116.2285 - mse: 74712.2031 - val_loss: 662.9263 - val_mae: 662.9263 - val_mse: 1419201.7500\n",
            "Epoch 44/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 120.0192 - mae: 120.0192 - mse: 80971.8750 - val_loss: 635.6360 - val_mae: 635.6360 - val_mse: 1345999.3750\n",
            "Epoch 45/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 110.9675 - mae: 110.9675 - mse: 74941.2891 - val_loss: 601.9926 - val_mae: 601.9926 - val_mse: 1244816.7500\n",
            "Epoch 46/100\n",
            "18/18 [==============================] - 0s 5ms/step - loss: 111.6012 - mae: 111.6012 - mse: 74565.6016 - val_loss: 593.3531 - val_mae: 593.3531 - val_mse: 1248764.3750\n",
            "Epoch 47/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 109.6161 - mae: 109.6161 - mse: 74190.7109 - val_loss: 637.0011 - val_mae: 637.0011 - val_mse: 1314970.2500\n",
            "Epoch 48/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 109.1368 - mae: 109.1368 - mse: 72792.5547 - val_loss: 598.4197 - val_mae: 598.4197 - val_mse: 1234610.1250\n",
            "Epoch 49/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 159.5531 - mae: 159.5531 - mse: 119669.8438INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 0s 29ms/step - loss: 112.4355 - mae: 112.4355 - mse: 73980.3359 - val_loss: 579.7042 - val_mae: 579.7042 - val_mse: 1208301.8750\n",
            "Epoch 50/100\n",
            "18/18 [==============================] - 0s 5ms/step - loss: 110.2108 - mae: 110.2108 - mse: 72939.7109 - val_loss: 639.2327 - val_mae: 639.2327 - val_mse: 1347994.1250\n",
            "Epoch 51/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 111.0958 - mae: 111.0958 - mse: 73281.9062 - val_loss: 597.3571 - val_mae: 597.3571 - val_mse: 1229791.1250\n",
            "Epoch 52/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 110.7351 - mae: 110.7351 - mse: 74529.6328 - val_loss: 580.7227 - val_mae: 580.7227 - val_mse: 1215477.1250\n",
            "Epoch 53/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 111.1785 - mae: 111.1785 - mse: 74149.4766 - val_loss: 648.3585 - val_mae: 648.3585 - val_mse: 1336951.1250\n",
            "Epoch 54/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 114.0831 - mae: 114.0831 - mse: 74815.0781 - val_loss: 593.2006 - val_mae: 593.2006 - val_mse: 1220661.2500\n",
            "Epoch 55/100\n",
            "17/18 [===========================>..] - ETA: 0s - loss: 112.0307 - mae: 112.0307 - mse: 76173.8828 INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 1s 31ms/step - loss: 110.4910 - mae: 110.4910 - mse: 74711.2500 - val_loss: 579.5067 - val_mae: 579.5067 - val_mse: 1211580.3750\n",
            "Epoch 56/100\n",
            "18/18 [==============================] - 0s 5ms/step - loss: 108.0488 - mae: 108.0488 - mse: 71844.1562 - val_loss: 807.3853 - val_mae: 807.3853 - val_mse: 1859293.0000\n",
            "Epoch 57/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 125.0614 - mae: 125.0614 - mse: 84025.2812 - val_loss: 674.1658 - val_mae: 674.1658 - val_mse: 1437687.2500\n",
            "Epoch 58/100\n",
            "18/18 [==============================] - 0s 5ms/step - loss: 115.4340 - mae: 115.4340 - mse: 73710.7500 - val_loss: 582.2697 - val_mae: 582.2697 - val_mse: 1202679.6250\n",
            "Epoch 59/100\n",
            "18/18 [==============================] - 0s 5ms/step - loss: 110.0881 - mae: 110.0881 - mse: 72845.1172 - val_loss: 606.7640 - val_mae: 606.7640 - val_mse: 1267544.0000\n",
            "Epoch 60/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 108.7156 - mae: 108.7156 - mse: 72675.7734 - val_loss: 602.3099 - val_mae: 602.3099 - val_mse: 1234689.8750\n",
            "Epoch 61/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 133.4817 - mae: 133.4817 - mse: 95984.3438INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 0s 28ms/step - loss: 108.1525 - mae: 108.1525 - mse: 72252.1250 - val_loss: 573.9990 - val_mae: 573.9990 - val_mse: 1189583.5000\n",
            "Epoch 62/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 107.3726 - mae: 107.3726 - mse: 70822.7891 - val_loss: 581.7015 - val_mae: 581.7015 - val_mse: 1214006.5000\n",
            "Epoch 63/100\n",
            "18/18 [==============================] - 0s 5ms/step - loss: 110.7667 - mae: 110.7667 - mse: 73178.4453 - val_loss: 637.5243 - val_mae: 637.5243 - val_mse: 1311201.6250\n",
            "Epoch 64/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 110.1539 - mae: 110.1539 - mse: 72588.2188 - val_loss: 586.6598 - val_mae: 586.6598 - val_mse: 1207116.3750\n",
            "Epoch 65/100\n",
            "17/18 [===========================>..] - ETA: 0s - loss: 109.1682 - mae: 109.1682 - mse: 72969.9609 INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 1s 32ms/step - loss: 108.2325 - mae: 108.2325 - mse: 71962.0781 - val_loss: 573.5618 - val_mae: 573.5618 - val_mse: 1188986.2500\n",
            "Epoch 66/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 112.6009 - mae: 112.6009 - mse: 72843.7344INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 0s 28ms/step - loss: 108.6826 - mae: 108.6826 - mse: 71026.3516 - val_loss: 572.2206 - val_mae: 572.2206 - val_mse: 1187818.2500\n",
            "Epoch 67/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 106.6371 - mae: 106.6371 - mse: 70166.2578 - val_loss: 646.6359 - val_mae: 646.6359 - val_mse: 1356393.0000\n",
            "Epoch 68/100\n",
            "18/18 [==============================] - 0s 5ms/step - loss: 114.1603 - mae: 114.1603 - mse: 71297.3203 - val_loss: 681.8572 - val_mae: 681.8572 - val_mse: 1455430.5000\n",
            "Epoch 69/100\n",
            "18/18 [==============================] - 0s 5ms/step - loss: 124.5514 - mae: 124.5514 - mse: 81238.9609 - val_loss: 655.9892 - val_mae: 655.9892 - val_mse: 1383324.7500\n",
            "Epoch 70/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 125.0234 - mae: 125.0234 - mse: 83058.2109 - val_loss: 601.0039 - val_mae: 601.0039 - val_mse: 1256963.6250\n",
            "Epoch 71/100\n",
            "18/18 [==============================] - 0s 5ms/step - loss: 110.3652 - mae: 110.3652 - mse: 71691.4922 - val_loss: 595.3957 - val_mae: 595.3957 - val_mse: 1222771.6250\n",
            "Epoch 72/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 107.9285 - mae: 107.9285 - mse: 71745.5391 - val_loss: 573.7084 - val_mae: 573.7084 - val_mse: 1192818.5000\n",
            "Epoch 73/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 109.5085 - mae: 109.5085 - mse: 71691.1484 - val_loss: 580.4183 - val_mae: 580.4183 - val_mse: 1211317.6250\n",
            "Epoch 74/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 108.7380 - mae: 108.7380 - mse: 72051.1094 - val_loss: 576.1216 - val_mae: 576.1216 - val_mse: 1199670.6250\n",
            "Epoch 75/100\n",
            "18/18 [==============================] - 0s 5ms/step - loss: 107.9404 - mae: 107.9404 - mse: 71043.1328 - val_loss: 591.1470 - val_mae: 591.1470 - val_mse: 1211107.3750\n",
            "Epoch 76/100\n",
            "18/18 [==============================] - 0s 5ms/step - loss: 109.4232 - mae: 109.4232 - mse: 72295.0234 - val_loss: 597.8611 - val_mae: 597.8611 - val_mse: 1243936.2500\n",
            "Epoch 77/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 85.1158 - mae: 85.1158 - mse: 65630.7578INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 0s 29ms/step - loss: 107.5879 - mae: 107.5879 - mse: 70333.8359 - val_loss: 571.9301 - val_mae: 571.9301 - val_mse: 1186830.0000\n",
            "Epoch 78/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 108.1598 - mae: 108.1598 - mse: 70200.0781 - val_loss: 575.2387 - val_mae: 575.2387 - val_mse: 1195318.0000\n",
            "Epoch 79/100\n",
            "18/18 [==============================] - 0s 5ms/step - loss: 107.9175 - mae: 107.9175 - mse: 72451.5859 - val_loss: 617.3083 - val_mae: 617.3083 - val_mse: 1285512.5000\n",
            "Epoch 80/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 108.9510 - mae: 108.9510 - mse: 72158.7812 - val_loss: 583.4841 - val_mae: 583.4841 - val_mse: 1194356.7500\n",
            "Epoch 81/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 96.3487 - mae: 96.3487 - mse: 44505.5117INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 1s 29ms/step - loss: 106.0505 - mae: 106.0505 - mse: 69690.9219 - val_loss: 570.0803 - val_mae: 570.0803 - val_mse: 1177237.8750\n",
            "Epoch 82/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 115.6827 - mae: 115.6827 - mse: 76903.4922 - val_loss: 575.7388 - val_mae: 575.7388 - val_mse: 1194662.6250\n",
            "Epoch 83/100\n",
            "18/18 [==============================] - 0s 5ms/step - loss: 110.9379 - mae: 110.9379 - mse: 72968.2969 - val_loss: 659.6579 - val_mae: 659.6579 - val_mse: 1388829.8750\n",
            "Epoch 84/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 111.4836 - mae: 111.4836 - mse: 73017.9375 - val_loss: 570.1957 - val_mae: 570.1957 - val_mse: 1177492.2500\n",
            "Epoch 85/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 107.5949 - mae: 107.5949 - mse: 70728.2266 - val_loss: 601.5938 - val_mae: 601.5938 - val_mse: 1228450.6250\n",
            "Epoch 86/100\n",
            "18/18 [==============================] - 0s 5ms/step - loss: 108.9425 - mae: 108.9425 - mse: 70066.7500 - val_loss: 592.8101 - val_mae: 592.8101 - val_mse: 1209680.7500\n",
            "Epoch 87/100\n",
            "18/18 [==============================] - 0s 5ms/step - loss: 105.7717 - mae: 105.7717 - mse: 68382.3516 - val_loss: 603.6174 - val_mae: 603.6174 - val_mse: 1252409.5000\n",
            "Epoch 88/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 135.4518 - mae: 135.4518 - mse: 130713.5312INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 0s 28ms/step - loss: 107.9217 - mae: 107.9217 - mse: 71668.0469 - val_loss: 569.0499 - val_mae: 569.0499 - val_mse: 1174588.3750\n",
            "Epoch 89/100\n",
            " 1/18 [>.............................] - ETA: 0s - loss: 101.8746 - mae: 101.8746 - mse: 57186.1758INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/assets\n",
            "18/18 [==============================] - 1s 30ms/step - loss: 106.0344 - mae: 106.0344 - mse: 68979.1484 - val_loss: 568.9510 - val_mae: 568.9510 - val_mse: 1171743.5000\n",
            "Epoch 90/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 105.4977 - mae: 105.4977 - mse: 67516.5000 - val_loss: 581.7678 - val_mae: 581.7678 - val_mse: 1190160.6250\n",
            "Epoch 91/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 108.8468 - mae: 108.8468 - mse: 71112.3359 - val_loss: 573.6025 - val_mae: 573.6025 - val_mse: 1185216.2500\n",
            "Epoch 92/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 110.8884 - mae: 110.8884 - mse: 73287.8125 - val_loss: 576.8243 - val_mae: 576.8243 - val_mse: 1181628.3750\n",
            "Epoch 93/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 113.8781 - mae: 113.8781 - mse: 75236.4375 - val_loss: 608.3022 - val_mae: 608.3022 - val_mse: 1264206.1250\n",
            "Epoch 94/100\n",
            "18/18 [==============================] - 0s 5ms/step - loss: 110.5763 - mae: 110.5763 - mse: 72590.8672 - val_loss: 601.6042 - val_mae: 601.6042 - val_mse: 1229105.0000\n",
            "Epoch 95/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 106.5906 - mae: 106.5906 - mse: 68612.3281 - val_loss: 570.3651 - val_mae: 570.3651 - val_mse: 1175759.3750\n",
            "Epoch 96/100\n",
            "18/18 [==============================] - 0s 5ms/step - loss: 116.9515 - mae: 116.9515 - mse: 75082.1172 - val_loss: 615.2584 - val_mae: 615.2584 - val_mse: 1277124.5000\n",
            "Epoch 97/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 108.0739 - mae: 108.0739 - mse: 71321.4219 - val_loss: 580.3076 - val_mae: 580.3076 - val_mse: 1202622.2500\n",
            "Epoch 98/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 108.7102 - mae: 108.7102 - mse: 70248.0703 - val_loss: 586.6510 - val_mae: 586.6510 - val_mse: 1200831.5000\n",
            "Epoch 99/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 109.0488 - mae: 109.0488 - mse: 70913.1641 - val_loss: 570.0628 - val_mae: 570.0628 - val_mse: 1177064.5000\n",
            "Epoch 100/100\n",
            "18/18 [==============================] - 0s 4ms/step - loss: 106.1845 - mae: 106.1845 - mse: 68681.7266 - val_loss: 585.9758 - val_mae: 585.9758 - val_mse: 1197802.1250\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7fc71036b0d0>"
            ]
          },
          "metadata": {},
          "execution_count": 44
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Evaluate model on test data\n",
        "model_1.evaluate(test_windows, test_labels)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7jJVJZe3utf1",
        "outputId": "4f5c3f5a-07aa-4c8b-f572-b9fbc350fe63"
      },
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "18/18 [==============================] - 0s 2ms/step - loss: 585.9758 - mae: 585.9758 - mse: 1197802.1250\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[585.9757690429688, 585.9757690429688, 1197802.125]"
            ]
          },
          "metadata": {},
          "execution_count": 45
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Load in saved best performing model_1 and evaluate it on test data\n",
        "model_1 = tf.keras.models.load_model(\"/content/drive/MyDrive/BTC Predictor/model_experiments/model_1_dense/\")\n",
        "model_1.evaluate(test_windows, test_labels)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "k6KKsS-i9I8i",
        "outputId": "ca3eb694-eae7-4189-c9f7-a1e9edd7fb67"
      },
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "18/18 [==============================] - 0s 2ms/step - loss: 568.9510 - mae: 568.9510 - mse: 1171743.6250\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[568.9510498046875, 568.9510498046875, 1171743.625]"
            ]
          },
          "metadata": {},
          "execution_count": 46
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Making forecasts with a model (on the test dataset)\n",
        "\n",
        "Writing a function to:\n",
        "\n",
        "1. Take in a train model\n",
        "2. Take in some input data (same kind of data the model was trained on)\n",
        "3. Passes the input data to the model's `predict()` method\n",
        "4. Returns the predictions"
      ],
      "metadata": {
        "id": "ycgebWJHDY7X"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def make_preds(model, input_data):\n",
        "  \"\"\"\n",
        "  Uses model to make predictions on input_data.\n",
        "  \"\"\"\n",
        "  forecast = model.predict(input_data)\n",
        "  return tf.squeeze(forecast) # return 1D array of predictions"
      ],
      "metadata": {
        "id": "Ubb20_y_DEHl"
      },
      "execution_count": 47,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Make predictions using model_1 on the test dataset and view results\n",
        "model_1_preds = make_preds(model_1, test_windows)\n",
        "len(model_1_preds), model_1_preds[:10]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "888KJbAADOPH",
        "outputId": "3b0b41f8-d510-4983-f9e8-cffa374b9c30"
      },
      "execution_count": 48,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(556, <tf.Tensor: shape=(10,), dtype=float32, numpy=\n",
              " array([8861.71 , 8769.887, 9015.711, 8795.519, 8723.811, 8730.111,\n",
              "        8691.952, 8502.054, 8460.961, 8516.548], dtype=float32)>)"
            ]
          },
          "metadata": {},
          "execution_count": 48
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "test_labels[:10]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DyiwUhNQFClI",
        "outputId": "9384b85c-9856-400b-8c75-5b20c84e196b"
      },
      "execution_count": 49,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[8798.042055],\n",
              "       [9081.186878],\n",
              "       [8711.534339],\n",
              "       [8760.892718],\n",
              "       [8749.520591],\n",
              "       [8656.970922],\n",
              "       [8500.643558],\n",
              "       [8469.260899],\n",
              "       [8537.339652],\n",
              "       [8205.806366]])"
            ]
          },
          "metadata": {},
          "execution_count": 49
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Comparing predicted values of model_1 with the test_values\n",
        "\n",
        "#But first we have to get test_labels and model_1_preds in the same shape\n",
        "\n",
        "test_labels.shape, model_1_preds.shape # hence we'll use tf.squeeze for test_labels"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "V_OhbgmQGQKm",
        "outputId": "95619d0a-6827-4174-abd7-c3f243da6c39"
      },
      "execution_count": 50,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((556, 1), TensorShape([556]))"
            ]
          },
          "metadata": {},
          "execution_count": 50
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model_1_results = evaluate_preds(y_true = tf.squeeze(test_labels), y_pred = model_1_preds)\n",
        "\n",
        "model_1_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "yqIXUhcCFV4A",
        "outputId": "8535c13e-b5df-476c-914b-fbe05d437929"
      },
      "execution_count": 51,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'mae': 568.95105,\n",
              " 'mape': 2.5448983,\n",
              " 'mase': 0.9994894,\n",
              " 'mse': 1171743.8,\n",
              " 'rmse': 1082.4711}"
            ]
          },
          "metadata": {},
          "execution_count": 51
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "naive_results # As we can see naive model (baseline model) beats model_1 "
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YGc06OIyFsB_",
        "outputId": "ba82c0f4-b98a-438e-83ce-a099086b2f00"
      },
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'mae': 567.9802,\n",
              " 'mape': 2.516525,\n",
              " 'mase': 0.99957,\n",
              " 'mse': 1147547.0,\n",
              " 'rmse': 1071.2362}"
            ]
          },
          "metadata": {},
          "execution_count": 52
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# PLotting model_1 predictions\n",
        "offset = 450\n",
        "plt.figure(figsize=(10,7))\n",
        "\n",
        "plot_time_series(timesteps = X_test[-len(test_windows):], \n",
        "                 values = test_labels[:, 0],\n",
        "                 start = offset,\n",
        "                 label = \"Test Data\")\n",
        "\n",
        "plot_time_series(timesteps = X_test[-len(test_windows):],\n",
        "                 values = model_1_preds,\n",
        "                 start = offset,\n",
        "                 format = \"-\",\n",
        "                 label = \"model_1_preds\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 446
        },
        "id": "7H6gXun2Gut2",
        "outputId": "961bd422-78c7-4289-ac60-ab9dd5d04b3f"
      },
      "execution_count": 53,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x504 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAnAAAAGtCAYAAACWW0nPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeXhU5dn48e+TlYSE7AlbEggh7GtQQFFAQVxfrIobIGJt1aqv/Nxq61KX2lctWpdad8UqVi1iVdxAakBlD4KsgQQIBMhCCAkJZD+/P56ZJJPMZD2TWXJ/rotryJkz5zxzMpm551nuWxmGgRBCCCGE8Bw+rm6AEEIIIYRoGwnghBBCCCE8jARwQgghhBAeRgI4IYQQQggPIwGcEEIIIYSHkQBOCCGEEMLDODWAU0qFK6WWKKV2K6V2KaUmKqUeVUodVkptsfy7uMH+f1BKZSqlMpRSMxpsv9CyLVMp9UCD7f2VUust2z9SSgU48/kIIYQQQrgD5cw8cEqpd4EfDMN40xJcBQMLgFLDMBY22nco8C/gTKA38B2QYrl7DzAdyAE2AtcZhrFTKfUxsNQwjA+VUq8CWw3DeMVpT0gIIYQQwg04rQdOKRUGnAu8BWAYRqVhGCeaechM4EPDMCoMw9gPZKKDuTOBTMMw9hmGUQl8CMxUSingPGCJ5fHvApc759kIIYQQQrgPPyceuz9QALyjlBoFpAN3We67Qyl1A7AJuMcwjCKgD7CuweNzLNsADjXaPh6IAk4YhlFtZ3+HoqOjjX79+rXrCdlTVlZG9+7dTTuep5PrYUuuRz25FuaQ62hLrkfbyTWz5e7XIz09/ZhhGDGNtzszgPMDxgJ3GoaxXin1AvAA8HfgCcCw3D4L3OTEdqCU+i3wW4C4uDgWLlzYwiNar7S0lJCQENOO5+nketiS61FProU55DrakuvRdnLNbLn79Zg6dWq2ve3ODOBygBzDMNZbfl4CPGAYRp51B6XUG8Ayy4+HgfgGj+9r2YaD7YVAuFLKz9IL13B/G4ZhvA68DjBu3DhjypQpHXhattLS0jDzeJ5OroctuR715FqYQ66jLbkebSfXzJanXg+nzYEzDCMXOKSUGmTZdD6wUynVq8FuvwK2W/7/OXCtUipQKdUfGAhsQC9aGGhZcRoAXAt8bujVF98DV1kePw/4zFnPRwghhBDCXTizBw7gTmCxJfDaB8wHXlRKjUYPoR4AbgEwDGOHZVXpTqAauN0wjBoApdQdwLeAL/C2YRg7LMf/PfChUurPwM9YFkwIIYQQQngzpwZwhmFsAcY12jy3mf2fBJ60s/0r4Cs72/ehV6kKIYQQQnQZUolBCCGEEMLDOHsIVQghhPA4tbW15OTkUFZW5uqmmC4sLIxdu3a5uhluw5XXw9/fn9jYWHr06NHmx0oAJ4QQQjRy7NgxlFIMGjQIHx/vGqw6efIkoaGhrm6G23DV9TAMg9OnT3P4sE6g0dYgzrtelUIIIYQJTpw4QVxcnNcFb8J9KKUIDg6mT58+5Ofnt/nx8soUQgghGqmpqcHf39/VzRBdQFBQEFVVVW1+nARwQgghhB265LYQztXe15kEcEIIIYQQHkYCOCGEEEIIDyMBnBBCCOHhlFLN/rvxxhvbfexHH32U4cOHt7jfokWL6s7n6+tLeHg448aN48EHH2zXJH2lFEuWLGlPk7sESSMihBBCeLijR4/W/X/ZsmX85je/sdkWFBTUKe0IDg4mKysLwzAoKSlh48aNPP3007zxxhusWrWKIUOGdEo7ugLpgRNCCOHdThe5ugVO17Nnz7p/4eHhTbatXr2a1NRUunXrxogRI3jwwQeprKyse/zSpUsZOXIkQUFBREZGMnnyZPLy8li0aBGPPfYYO3bsqOtdW7RokcN2KKXo2bMnvXr1YtCgQcyZM4e1a9cSHh7OrbfeWrffxo0bueCCC4iOjqZHjx5MmjSJtWvX1t3fr18/AGbNmoVSqu7nrKwsZs6cSc+ePenevTtjx45l2bJl5l1IDyIBnBBCCO9UXQFf3Q9P94M937qkCenZRbz8fSbp2a4LIr/99ltmz57NHXfcwY4dO3j55ZdZsmQJf/zjHwHIzc3l2muvZd68eezatYvVq1czd64uW37NNddwzz33MGjQII4ePcrRo0e55ppr2nT+kJAQbr31VlavXk1BQQGgk+fOnTuXH374gQ0bNjB69GguvvhiCgsLAR3gAbzxxhscPXq07ufS0lIuuugiVqxYwdatW7nyyiu54oor2L17tynXypPIEKoQQgivkZ5dxLp9hUyOKmb4mgWQ+wv4BsAvH0PKjE5vy+w311FZXUuAnw+Lb55AamJEp7YB4Mknn+S+++5j/vz5AMTGxvL0008zZ84c/vrXv3LkyBGqqqq46qqrSExMBLCZ8xYSEoKfnx89e/ZsdxuGDh0KwP79+4mJieG8886zuf+ll17ik08+4euvv2bOnDnExMQAEB4ebnPeUaNGMWrUqLqfH3zwQb744guWLFnCQw891O72eSLpgRNCCOH2WtOTZQ2YMr97i/6fXEx10UG47kMYeY3ugauu6MQWw7p9hVRW11JrQFV1Lev2FXbq+a3S09N58sknCQkJISQkhF69enH99ddTVlZGbm4uo0aNYtq0aQwfPpwrr7ySV155pa6nzCyGYQD1Oc/y8/O55ZZbSElJISwsjNDQUPLz8zl48GCzxykrK+P+++9n6NChREREEBISwqZNm1p8nDeSHjghhBBurbU9WRsyc3mMV7nGP42NtYPYOfo55g2aBMoHfn4P9q+GgdM7rd0TkqII8POhqroWfz8fJiRFddq5G6qtreVPf/oTs2bNAvQwZEhICAAxMTH4+vqyfPly1q1bx/Lly3nrrbf4wx/+wKpVq2x6uzpi586dNnPZ5s2bR15eHn/729/o168fgYGBnH/++Tbz8uy59957+eabb1i4cCEDBw4kODiYG264ocXHeSMJ4IQQQrg1ez1ZTQK4yjKu3/d7wnxX8Y/qmbysruafQ4fp+/pPhoAQ2PVFpwZwqYkRLL55Auv2FTIhKcolw6cAY8eOZffu3SQnJwP2i7crpZg4cSITJ07kkUceYdiwYXz00UeMGjWKgIAAampq2n3+0tJSXn31VSZPnlw3NPrjjz/y4osvcskllwCQl5dns2oWwN/fv8l5f/zxR2644QauvPJKAMrLy8nKyiIlJaXd7fNUEsAJIYRway32ZJ0ugg+uIezIRg6c/RSG33T+2TBg8u8GAy+AjK+g9m/g49tpbU9NjHBZ4Gb1yCOPcOmll5KYmMjVV19NRUUF+/fvZ8OGDTzzzDOsW7eO7777jhkzZhAXF8fPP//MoUOH6uat9evXj+zsbDZv3kxCQgKhoaEEBgbaPZdhGOTm5gJQXFxcl0akuLiYzz//vG6/lJQU3n//fcaPH183LBoQEGBzrH79+rFy5UomT55MYGAgERERpKSk8OmnnzJz5kz8/f157LHHKC8vd9KVc28SwAkhhHBrjnqy0rOLWLUnj/FbZhB0cj/MWkS/oTO53d5BhlwKO5bCofWQeFantt/VZsyYwZdffskTTzzBwoUL8fPzIyUlpS65b1hYGD/99BMvvfQSJ06cID4+nocffpg5c+YAcOWVV7J06VLOP/98Tpw4wTvvvOMwMfCpU6fo1asXSilCQ0MZMGAAl112GQsWLCA2NrZuv7fffpvf/va3pKam0rt3bx599NEm8+6effZZ7r77buLj4+nTpw8HDhzgueee49e//jXnnHMOERERLFiwoMsGcMo6sbCrGDdunLFp0ybTjpeWlsaUKVNMO56nk+thS65HPbkW5pDrqKVnF/HAm//hLfUkUaqEwxe+RcrEyxw/oLwE/joAzvgNXPiXFo+/a9cur006a28ItStzh+vR3OtNKZVuGMa4xttlFaoQQgiPs25fIQvVi4Sq08ypepAV5S0EW916QNIU2P0FdLGOC+GdJIATQgjhcSb0C2eIyuajmqns8k1p3QrPIZfBiYOQu835DRTCySSAE0II4XFSw0oIUDVExMS3PkHuoIt1SpFdXzi/gUI4mQRwQgghPM+xTAAG9kto/SrP7tGQMBF2d83amcK7SAAnhBDC8xTqAO50UJ+2PW7wpZC/EwqznNAoITqPBHBCCCE8T+Fe6BZOlX8bVw8O1oljpRdOeDoJ4IQQQniewkyIHgiW2pqtFpEIvUbJPDjh8SSAE0II4XmOZULUwPY9dvBlkLMRSo6Y2yYhOpEEcEIIITxLRSmcPAJRA9r3+KEz9e3Oz8xrkxCdTAI4IYQQnuW4ZQFCdOt64NKzi3j5+0zSs4v0hpgUiBsB25c6qYHe74477mhzRRClFEuWLHFOg1xo06ZNKKU4cOBAp55XAjghhBCe5dhefRuV3OKu6dlFzH5zHc8uz2D2m+vqg7jhv4KcDTqxr3A7Tz75JGeffTbdu3dHtXWeYxchAZwQQgjPUpgFKIhManHXdfsKqayupdaAqupa1u0r1HcMu0Lf7vjUee0U7VZRUcEVV1zBggULOuV8tbW11NTUdMq5zCIBnBBCCM9SuBfC48E/qMVdJyRFEeDng68Cfz+f+pJbkf2h91ivG0adMmUKt912G/fccw+RkZHExMTwwgsvUFFRwe233054eDhDhw7lvffeq3vMtm3bmDZtGkFBQURGRnLjjTdSXFxcd39NTQ333nsvERERREREsGDBgibBjmEYPPPMMwwYMICgoCBGjBjB+++/3+7n8fjjj3PPPfcwZsyYNj82LS0NpRTLli1j9OjRdOvWjdTUVNLT0+v2WbRoESEhIXz11VeMHz+egIAAdu3aRWVlJb///e/p27cvwcHBnHHGGXz77bc2x//mm28YPHgw3bp145xzzmHPnj029xcXFzN37lxiY2Pp1q0bSUlJPP/88+27EM2QAE4IIYRnKcxs1fApQGpiBItvnsDdFwxqWnJr+BVwdIvXJfVdvHgxoaGhrF+/ngceeIAFCxZw+eWXk5KSwqZNm7juuuu4+eabOXr0KGVlZcyYMYOQkBA2bNjAp59+ypo1a7jpppvqjvfss8/yxhtv8Nprr7F27VpqampYvHixzTkfeugh3nrrLV5++WV27tzJH/7wB2655Ra+/PLLzn76de69916efvppNm3aRFJSEpdeeimnTp2qu7+8vJwnnniCF154gZ07d5KYmMj8+fNZtWoVH3zwAdu3b2fevHlcdtllbN26FYBDhw5x+eWXM336dLZs2cKdd97J/fffb3Pehx56iG3btrFs2TIyMjJ4++236dOnjQmnW8HP9CMKIYQQzmIYOoXI6Otb/ZDUxAj75baG/QqWPwQ7lsK59zV/kK8fgNxtbWxsB/UcARc91eaHDRs2jEcffRSAu+++m6eeegp/f3/uuusuAB544AGef/55fvrpJ4qKiigrK+O9994jNFQnRX799deZOnUqmZmZJCcn8/zzz3P//fdz9dVXA/DCCy/Y9EqVlZXx3HPPsXz5cs455xwA+vfvz4YNG3j55Ze55JJLOnIV2u3hhx9mxowZALzzzjv07duXDz74gJtvvhnQPYt///vfSUlJITQ0lKysLP71r39x4MABEhISAL1Y47vvvuO1117jH//4B6+88goJCQm8+OKLKKUYPHgwe/bs4eGHH647b3Z2NmPHjuXMM88EIDEx0SnPTwI4IYQQnqM0DypPtroHrllhfSF+Amz/tOUAzoOMHDmy7v9KKWJjYxkxYkTdNn9/fyIiIsjPzyczM5ORI0fWBW8AZ511Fj4+PuzcuZOYmBiOHj3KxIkT6+738fFh/PjxHDp0CICdO3dSXl7OhRdeaLPgoKqqin79+jnxmTavYZtDQkIYMWIEO3furNvm5+fH6NGj63rlNm/ejGEYDB061OY4FRUVnHfeeQDs2rWLCRMm2DzPhucBuO2227jqqqtIT09n+vTpXHbZZUyePNn05ycBnBBCCM9hqYFKtAkBHOhh1K/vh/zdEDvY8X7t6AlzFX9/f5uflVJ2t9XW1jZ7nNau/rQe54svvqjruXLUFncSGBiIr69v3c+1tbUopdi4cWOTdgcFtTzf0uqiiy4iOzubr7/+mpUrV3LJJZcwa9Ys3nnnHdPaDjIHTgghhCepSyHSzioMjQ29HJSPHkbtgoYMGcK2bds4efJk3bY1a9ZQW1vLkCFDCAsLo1evXqxbt67ufsMw2LBhQ93PQ4cOJTAwkOzsbJKTk23+OWv4sDUatrmsrIzt27czZMgQh/uPGTMGwzDIzc1t8jysc9iGDBnC+vXrMQzD7nmsoqOjmTt3LosWLeKtt97i3XffpaKiwsRnJwGcEEIIT1KYCX5B0MOkSeGhcZB4Nmz/RM+v62Jmz55NcHAwN9xwA9u2bWP16tXccsstXHHFFSQn617Ou+66i2eeeYYlS5aQkZHBggULOHr0aN0xQkNDuffee7n33nt5++23yczMZMuWLbz66qu8/vrr7WrXwYMH2bJlS11y3C1btrBlyxZKS0tbfYw///nPrFixgh07dnDTTTcREBDA9dc7njuZkpLC7NmzufHGG1myZAn79u1j06ZNLFy4kKVLdYB/6623cuDAARYsWEBGRgZLlizh1VdftTnOI488wn/+8x/27t3Lrl27WLp0KUlJSQQGBrb9QjRDAjghhBCeozBTl9DyMfHja/gV+ridvUjBDQQHB/Ptt99SUlLCmWeeycyZM5k4cSJvv/123T733HMP8+fP5+abb2b8+PHU1tYye/Zsm+M88cQTPProoyxcuJBhw4Yxffp0PvnkE/r379+udj3yyCOMGTOG++7TcxPHjBnDmDFj2LRpU6uP8dRTT3HPPfcwduxY9u7dy7Jly+jevXuzj3nnnXeYP38+999/P4MHD+bSSy9l9erVdT2JCQkJLF26lG+++YZRo0bxt7/9jaeesh1eDwwM5MEHH2TUqFGcffbZnDx5ki+++KKNV6Blyuhi3zjGjRtntOUF0JK0tLQ2lxPxZnI9bMn1qCfXwhxd/jq+OBZ6Doer/wmYdD3KCmHhQDj7f2Hao4CerN7ccJsnO3nypM2iBW+TlpbG1KlTKSgoIDo6usX93eF6NPd6U0qlG4YxrvF26YETQgjhGWqqoOiAefPfrLpHQdIUr0vqK7ybBHBCCFM1KRwuhFmKDoBR0+oi9m3S/xw4kQ0VJ1veV3TIX/7yF0JCQuz+u+iii1p8/K233urw8bfeemsnPAP3IGlEhBCmsRYOr6yuJcDPp2nmeyE6wppCxIwccI2Fxevb4sPNpxMRHXbrrbfWJQVurDXpOh5//HHuvfdeu/f16NGD2NhYusL0MAnghBCmsVc4XAI4YZq6FCIDOnSY9Owi1u0rZEJSVP3r07qqtSRHAjgni4yMJDIyst2Pj42NJTY21sQWeSYJ4IQQprEWDq+qrrUtHA46RUPRATjwE2T/BNlrYOhMmP6Yy9orPEzhXgiOhqD2fylw2Esc1lfvUJxjUmOFcC4J4IQQprEWDm/Su7HjUyasuxdWHdM/B1sCu8yVEsCJ1ivM6vD8N4e9xKG9dELf4sN1+xqG0epqBEK0V3uHeyWAE0KYym7h8K0f4VNbDRcv1ElTYwbD1/fBtn+7ppHCMx3bCykzOnQIh73Evn46iLP0wPn6+lJVVUVAQEBHWy1Es06fPt2ukmMSwAkhnK8wk+KwIcSc+Zv6bWHxUF4M5SXQrYfr2iY8Q3kxlOV3eAGDw15i0PPgSnQAFx4eTl5eHn369MHHzKTBQlgYhsHp06c5fPgwcXFxbX68UwM4pVQ48CYwHDCAm4AM4COgH3AAuNowjCKl+6lfAC4GTgE3Goax2XKcecBDlsP+2TCMdy3bU4FFQBDwFXCX0RWWngjhSWqqMY7v55eQ4YRnF9V/YIZbV/0dgm7DXNc+4Rnqith3PIWI3V5i0PPgjm7Vp4mOJicnh4yMjA6fz92Ul5fTrVs3VzfDbbjyevj7+xMXF0ePHm3/EuvsHrgXgG8Mw7hKKRUABAN/BFYahvGUUuoB4AHg98BFwEDLv/HAK8B4pVQk8CdgHDoITFdKfW4YRpFln98A69EB3IXA105+TkKINti+cxvDjWq+OR7DF2+uazBp3BrA5UCcBHCiBcecmELEKqwPZHwFhoGPjw8JCQnOO5cLpaWlMWbMGFc3w2146vVwWr+wUioMOBd4C8AwjErDME4AM4F3Lbu9C1xu+f9M4J+Gtg4IV0r1AmYAKwzDOG4J2lYAF1ru62EYxjpLr9s/GxxLCOEm9mf8AsC+2l51k8aB+gDuxEEXtUx4lMJMvcggon21NVulR1+oLodThc47hxAmcebAfn+gAHhHKfWzUupNpVR3IM4wjKOWfXIB68BvH+BQg8fnWLY1tz3HznYhhBsZHaxXnmYbvWwnjYfEgY+/HkIVojn5u2H7JxCZBH5OXFRQl0pEXpPC/TlzCNUPGAvcaRjGeqXUC+jh0jqGYRhKKafPWVNK/Rb4LUBcXBxpaWmmHbu0tNTU43k6uR625HrAwJyNVPh2Z2JCCKN6BXBy/1bS9uv7xgdEUbJ3M7v801zaRk/iba+pzKIadh+vYXCkL8kRvrbbInw4p/y/JGe+QY1vEDuH3sOJRs/dzOsRcjKXccD2n77lWEyxKcd0R972GuooT70ezgzgcoAcwzDWW35egg7g8pRSvQzDOGoZBs233H8YiG/w+L6WbYeBKY22p1m297WzfxOGYbwOvA4wbtw4Y8qUKfZ2a5e0tDTMPJ6nk+thS64HcPBvEDeIK1NCm16LAwMJqq4grqtfozbwptdUenYRC1dak+rWsPjmCQAsXLmObtUlTAt4i0FqPSRNwfdXrzE6tGeTY5h6PcqGQ/o9DI8PhwkmHdMNedNryAyeej2cNoRqGEYucEgpNciy6XxgJ/A5MM+ybR7wmeX/nwM3KG0CUGwZav0WuEApFaGUigAuAL613FeilJpgWcF6Q4NjCSHcReE+xxPPwxNkuKoLs5dUd92+QgbU7GNZwB85n02sSboL5nwKdoI30wVHgV+3ulQiQrgzZ69CvRNYbFmBug+Yjw4aP1ZK/RrIBqwVbb9CpxDJRKcRmQ9gGMZxpdQTwEbLfo8bhnHc8v/fUZ9G5GtkBaoQ7qXqtA7QoubYvz+sL5zMhepK585tEm7JXlJdv/JCrvBfCMD1tY/xwOQ50Fl52JTSueCK7Q7mCOFWnBrAGYaxBZ3+o7Hz7exrALc7OM7bwNt2tm9C55gTQrij4/sBQxcft7ewLyxe319yGCKduLpQuKUmSXXje8D7N1LrV8q/R73NA6Mm2c/X5kxhfaQeqvAIUolBCOE81uSrUQOg0M6k8IbJfCWA65JskuqufAL2peHzPy9xzdjLXNOgsHjYl+aacwvRBlIfRAjhPMez9G3kAPv31+WCk3lwXUV6dhEvf59JenaR7R0ZX8MPC2HMHBh7g2saB3oI9eRRqKl2XRuEaAXpgRNCOE9hps735qjWaQ9L6kYZsuoS0rOLmP2mddWpT31VjuP7Yekt0HMkXLzQtY0M6wNGrQ7iwuNb3l8IF5EeOCGE8xRmNV/6yL+bDvCKpRpDV2Bv1SlVp+HjuaCAa94D/yDXNtKazLdEFjII9yYBnBDCeQozdfb85oTFyxBqF2FddeqrqK/KsekdyN0Gv3odIvq5uom6nBZIr7BwezKEKoRwjvJiKCtoufh4WF/9AS7aLnc7J/9zDwEF26gOT6J778EQnaKvuX8wVJyEimIoL4GqUzDqOr2gxEWarjoNg8/ehPjxMOhCl7XLRpgM6wvPIAGcEMI5Ci0LGFoK4MLj9QT22trOy/fl6cqL4fu/YGx4g+raYL6oHU/isQLGla8hcNu/HT+uMAtmvePw7vTsovrgyknpO2xWnWZ9rxe6TP69U87VLoGh0C1MAjjh9iSAE0I4R2sDuLAEqKnQvXWhcc5vl4ewG0wZBnG5/4WXfgNlBWzvfSU37L+AIiMEXwV3jxvExIRgHlv0OaqmggrfEJ66fhKjBsTD8odgywe6Vy4w1O757C4wcKZNb+nqB0NnOvc8bRUWL3PgXKAzvkB4E/m6K4RwjsJMQLU8r6lhLjgB1AdTzy7PYPab6+pSbhz85gWG7H6B0uDe8NvvqZyxkNN+PWzmlK09eIpt1fFsqU1mT3VPfsz1hYBgGDELqk/D7q/sntPuAgMAw4CKUvOfZPFh3ZYxc/RiFnfSo4+8HjuZo9e8cEwCOCGEcxzP0sFZSx/O1lV/Dj4wHeYNa0F7H+cO7AVT6dlF5K37kF21CZyRez/pVf3q5pTdfcGguh4zuwsFQM8zC4uH7UvsntPh41Y8DAtT4HC6uU9y87s6XUfqfHOPa4awvlJOq5M5/AIhHJIhVCGEcxRmtjx8Cs0m823vsJ7Dxx3LBL9At8/vZa9G6OY9B5hPBq/WXkZljf7As84na3hNmiwUsN7n4wPDr4Q1L0HZMegebXNOu487+gusfVnv8MG18JuVEJ7Q8SdYUwXp78LA6e5ZgSOsD5w+DpWndO+lcDp7r3nRPOmBE0KYzzBazgFnFRQOgT3s9sC19lt54942u4/7+X14ZSK8cR4UZXfo6TmbvZ618wJ24qdqWV0zqsUPuNTECG6fmtw02B1xFRg1sPM/LT+utha+vBuCImH+N1BdAYuv1gsoOmr3l1CaC+N+3fFjOUMPyQXX2ey95ps4fQJ+fB7euUR6SJEeOCGEM5QVQEWJ4xJajYX1tbvqrzXfyu31tjV8XJBfLVflvwSrFkHi2ZC3HRbPgl9/C0HuO1G6cc/agKI1VAf0oF/fofx+ejsXGMQNh5jBsG0JnHFz8/v+/B7kbITLX4GE8XDNP+H9K+HjeTD73+Dr3/bzW216Sy9eGTi9/cdwprAGueCiB7q2LV1I49d8nRMHYd2reti90jIfM2ula0uuuQEJ4IQQ5mvtClQrB8l87Q7rVVfCLx/q4S0MTu/NZ66RxymfQHbW9GNDZiK3nT+UxTdPYEtGJrP2PUyPXWthwu0w/XE4tA7e+xV8OAfmLtVDqu7OMCDzO/wGns/FMUHtX6GnlO6F+++f9fV2NJRcVgjf/QkSztK545fdxP4AACAASURBVACSpsClz8Pnd+ieucte1Mdrq4IM2L8azv8T+Pi273k4m+SCcw+1NbBsAfy8WP88/EqY+Dt4+yLI2+natrkBCeCEEOYrzNS3rU0aGx6vAys7mnwr370MPr+z7sdJwKQG72S1Pz0BWSNI7TOW1L3L4WQeXP4qjLYEIv0mwcx/wNKb4bM74IrX2xeIdKbcbXrIceAFcKKDxxpuCeC2f0J6/Dz7aRtWPqqT/17yrO21GTsXivbDD8/qhMFn3dnk8A3ZTQux6W3w8Ycxczv4RJwotDegZAjV1bK+h83/1D1t595f/4UjdjDk73Bt29yABHBCCPMVZuoP6dZOeA/rq+dWlZc4LnxvdWgD+AXBgm26B0f58HNOMb9kHmRScA4DKnfrFZNbP4Ru4TD/a+ibanuMkbPgRDb89wndxvMfbt/z7Cx7l+vb5GmwqYM9D5H9oc84Tm3+kNnfDmq60OPQBv2hedadEDe0aRA29SEdUK5eCBN+57AXze5Ckjhf2PIvGHY5hMR07Hk4k1+ApUavpBJxqc2LqAqM5M3ut3FmcQip4ZbtscNgzzcubZo7kABOCGG+wkwdKLR2iMy6ErU4B7oNbXbX0n1rKek+mKOFfnW9OmMGhjNmYKLtjrU1gHJc3eGce/Tcmh8WQs8ROqhwV5nfQa9REBILmDB0NGIWwd/8noSag+wx+tYt9EjtGwrL7tY9UJMfcLyad+Q1Oqg8uhX6jLV7iiYLSbKOkbr+CT2HacLvOv4cnE1SibhWaT7G7q95r2YGf/1uPwFp2fWvv7ihsOV9KM23/E10TbIKVQjRfqUF8NntkLvddvvxfa2f/wb1PXUt9Hhs3pdLQP52vijs23KyTx/f5ktzKQWXPAfB0bB3Revb2tlOF8Gh9Xr41CzDfoWhfLjcby2+CgL84EKf9fDq2ZC3DS56CgJDHK8C7n+uvt2XVnfIxiuBG+eV+5/y/8DOz2Daow6DPrcS1keGUF1pywcoo5oPqqc0ff3FWr7k5XXtYVQJ4IQQ7Ze+SKfneHOaLtMEOv1EYVbbiqbX5YI72Oxu+35ZQ4CqZnNtcruSfTZJ7uvrB5FJejjVXWV9rxPeJpu4YjM0DtX/XH4dls6LY3NJj3mCAd//Tp9n1qK60lYOk/uGxOoVrZYAzl4W/YZpIT671If4jf8HQy5rcd6c2wiL1z3ChuHqlrjGD89BhouGKQ0DNv+Tk3FnkOMb3/T1FzdM3+Z37YUMMoQqhGgfw4BtH0PvMRAQAv+5TfcUTbxD1zZtSw9cSJyeM9fCqr/xAXp161Yjuc3JPh0OB0YkwsH1rW9rZ8v8Tqc76TvO3OOOmEXgZ7dzyY67dRD7q9f1CtUGw94OkwKDXpW64Q2oOm23p64uyXBkJbz2v7qk2syX3X/BiFWPPlB1SveABke6ujWdK28HrHwMIvrrnt/merKdIfsnOJ5F6OX3sTjCzusvJFb3nHfxlagSwAkhWs1mQrt/Nhzbo1NLjJkL3/8Zfvxbfa3N1uaAA/0BEdZy/cn4sh1UhPRh7uQJbS547SjIIDwRtn+iqwN0JLdZB9ldsVlbq4d3B5xnfsqNoZfrdB79ztGpQnztfxw4zM2VNBXW/h0OrmVC0hj7+fpqqmHJTXqBytyl0C3M3OfgTA1TiXh5ANfktbfmJX1H0X7YvwoGTO3kBr0LgWEwdCapAcH2X39xQ7v8SlQJ4IQQrdK4B2vVyO+I8/HXw22+fnpuU/x4+PQW/YC2JkB1kAvOxqGNBCaO5/apbejds3CYFDgiUQ8dFue4rKyTw97B3F+gLN+U4dMmH9KBITqFSnslTtS9pvvSSJ1+nv2euv8+Dtk/6t4967CXpwhrUI2h10jXtsWJGr/2Pr4ukZHb/g2pN+o5i+mLOjeAO12kzzt2bvNlzGKH6bbV1rhvPkEnkwBOCNEqDXuwaqqr6b73Mz280rB3YtBFcMsPutpBaM+2nSAs3mZSfBMlR6AkB/re3q72OxwODLesXj2R7bIAzmHvYKZlcUXytA4dv701ZZsV0F0H7JbfWZOeusIs+OlFGDsPRl3TsXO5Qo8G1Ri8WOPXXs2al/X0iEl3g3932PBa5672/OVjPQWjpSoLcUOh+jQUHWjbfFsvIosYhBCt0nBC+1l+uwmpLNBzphqLSITBl7T9BOHxcPKorrRgT84mfRt/ZtuPbWG3RmiENYBrfgGFMzlcLLB3hZ5j2MGcaa2tKdtmSVN0wfsyO8db/xr4+MHUP5pzrs7WPQZ8A7w+gGv42ov0O83I3E9h+BX67yL1Rqit1guVOoNh6OHTXqN12pzmxFp6dLvwSlQJ4IQQrdJwVeGzg/dAQCibgybarursiLB4wHCcuiFng/5A7Tmi4+dqqEdfUL4uLXBvt5D3qeO6FqkJw6cOA8SOSpoCGHBgte320yf0h/6Iq9reE+sufHygR2+vSyWSnl3EsqzKur/Zhq+9T8/MwLe6DM76X71zTAokTtI1SGtrm67idnD8dr8nHN6s57Wlzmt539jBgOrSK1FlCFV0DWXHoHu0q1vh8VITI0jtHQQLv+VYwgVcv2iLecNyDQuI2xvKzNmkv5WbXbvU109PWHdxKpEmQ5D7V+m5eSYUfG92NWlH9B4DgT30MOqwX9Vv//k9qCqD8beacx5XCYt3ac+s2axD6RVVtSw7sK7ubzY1MYLUPsHw/Lt6cUrDOX/j5sMnv2bPui+Y/XVAs3/vHR6q37wI/IN1ubeWBHTXK5ulB04I79HkG+CRLfDXZDi00bUN8xZ7l0NFMWuCppo7LNdcMt+aKjjyM/Rt//Bp8+dOdGkPnF3Za/QcpN5jTDmc3eHjjvL106tYs76v31ZTDetfh8Szofdo887lCr1H616hU8dd3RJTWIfSDez8zf7yEZTmwdl32T5oyGUQHIXv5kUt/r23Zai+yfv0sUzY+pEuWN9SOT2ruGFdugdOAjjhVewlFCX7J8CAw5tc3TzvsO1j6B5Ln7EXmjss18OStsFeIJW7DarLzc+FZhWR6PIeuCay1+j5fi5MbdIqSVP0tTu+X/+c8SUUH4QJt7myVeYYeQ3UVsGOpa5uiSmsQ+k+NPqbra3VC056jrQMizfgFwijryepcBV9/Iqb/Xtv7VB9k/fpA4Xwxf+Cfzc476HWP6HYobrqS9Xp1j/Gi0gAJ7yK3W+AR37Wdxbsdm3jPIzduSynT8Ceb2H4FaT2j2k6b6sj/LtB77F6+K2yzPY+6wKGvmd07ByOhCfq3gd3+SA4XaSHhhLPdnVLWpY0Rd9aVxCv/Ye+noMudlGDTNRzhJ4sv/UjV7fEFNah9CsG+tv+ze75Ggr36t43e4mWU+ejjBo+PCOr/u89sgK+exRenqBXHONgLqcdjd+nS356S3/RvuDPbZszGTdMTzPoou/tMgdOeBW7ub62bdZ35nfNP/L2cDiXZdcXUFMJI64Gmkny2l4z/gLvXKjL+Jz/cP32nI0Q2qt+npzZwhusRI0Z5JxztMXB9YChc625u+iBuvd0X5qeO3VoHcz4P4/LzWU3kTLAyKvhuz+1vTycm0pNjODkgADb5/jz+3oxz9DL7T8oagD0P5c+WR9z+5yrYe0jsPVD/V4AOiG05dq05j2h4ft0H78TnHvgRV1fd8zctj0Za27BvJ2mTTXwJNIDJ7xKk2+AsQqOZ+lVhgW7u25dwzZyOJdl28e67JKzipEnTtTB4ZoX9dCIVc4GPXzqrDJM1lQi7jIPLvsnveK2T6qrW9IypXQv3P5VuvctIBTGzHF1q9rE7tQLqxGzAKXzk3mr4/v0fD8H1TgASJ2vh8ZfPlNfizFz4c7NetHBsT1tOl3d+/T0FD7v/ym+RpWu6NLWv+/IJPDr1mXnwUkAJ7yOzWTto1v0xoEXQPkJPUwmWmR3LktZIez/Qa8Qc1IglZ5dxKLu86lRfvCtZS5MaYFO1umsBQxgm8zXHWSv0cGbf5CrW9I6SVP0sO/2JTp4a+0kdDfR7OT7sD66d+iXj7zzC6Bh6AooYfHN7zf4Ut1Dd869sGA7XPqc7nWLSoaCjDafNjUxgtvjthN+cIXOFdie3k0fX91j3kVXosoQqvBu1vlvo67V8zwKdntuXqpOZDftxP7VOHNYr+Gw7TH/mdyb8S9dyL2mSu/grPlvACFx4BuoA0VXqyjVXzwarwZ0Z/0nW/6jYPxvXdqU9nBYZs1q1LXwn9vg0AbSjRTz07G40ukinfIlvIUAzi8Arn636faYQXBwXbMPzVq1mO6bXsG313BiBk3Qw52hveCr+3TS3gntq64CQOwwqvas4PXvM73nd9JKEsAJ73Z4s84VlGAJOgoymq6yEnY1mcti/ZYdM9gp52vYC/JW1YXcHP4T4V8/AIMu1Bn9nZmSwsdHpzFxh5xfORt19vuEs1zdEoeazBcLjdM9pOHxeljLw7SYJ2/IZbDsbgp+epfZOy8ztySZq1nT9rR3fmn0INj2b/3FIzCkyd3p2UUc/+87TCKDypJ9sOdf9XcqX5iztPmh2xYcCuhP/OkC3l6+kZf8wr3jd9JKEsAJ73Zki547FRILQRGQv8vVLfJcBbt10tbQXk45fMNeEMMvkIKzHyP8u/mwPluvBnT2cKK7pBI5uBaUT4dKhjmTwwUuN37pvDmKnaDZyfeBoTDkUkJ3fQHVF1Br+NvWrPVk1lJhjYZQHS7qaCwmRd8W7rW7kGDdvkIu4wAra8dyV/WdPHZOd+bEW7IDxA23TRrcDj+X9yIeSFGH2FDdwzt+J60kc+CE9yo7pifd9h6jP1hiBrdrroawKMjQwyUmfUg3TlPSeAHKwElXwMAZeqWbM4dPrdwlmW/2Gp2Py03nkTmcL+YX4P456zpi5LV0qy5hmt9W80uSuZKdAK7ZRR2NRVtWbRfYX8hwVl9/ElQBu40E/P18GTJstC6xNuNJGH1dh5ufOFR/0Rnic8h7fietJD1wwntZ579ZV0zGDIKdn+lJux7cU+Ay+btg0EWmHMpRL06TXpAL/w8Op+tFKM4WkagXupQXQ7cw0w7b6p4MgOoKPYQ67temnd9sLc4X8yI2v7ukKdA9lieidjCk/2y9rXeQXnm7/lWY/phtOTFPceKgXsnZoNSgvSDd4Ws3MkkPhR6z/+V4TICuJTt49FksPsP84c1Rg1OoDozgyp4lXHKR5fj5u+D7J3WC4us+MPV87kQCOOG9Dm8GlK6fCRAzBE4vgrICPaQqWq/sGJw6Ztr8t1Z/QEQNgPsyOyfgDm+QSqSDwzpWba4NeeRnXXHCjfO/Oa2uqpux+7sbMYuIDa9z+9VhkPE5fPK0LnavfOGXf3tmAFeco+e/Nfgba1OQ7hegaxc7Gt3I3Q7ApRfMgB5OeK0ohV+v4QyryoGwk/CfB2Hrv3SCX9Bfisyun+wmJIAT3uvIzxCdouevQH2C1vxdEsC1lckLGNr0AdFZvaURDVKJmBTAOQxUTx3XvXyNk91m/6RvE9w3gAMnJHB2Q3Z/dyOvhnUvw0tjoaJED+1f/opOn7LjM6it8bgExhQfarKAoc1BevQgx7ng8rZBUKTT5s4CuqTWprfgpVRAwYTf6WwDyx/SAaoXJGC2RwI44Z0MA45shqSp9dtih+jbggxImmz/ccI+a6kak6oUuGUvTrj5yXztBqo11fqDJvEsuPo9vQLWKnutDpIbDGcJ17D7u+s1QAfXFaVw3oOQcqH+glGaD5v/Cbm/eF5FgOIcu1MU2hSkx6TA3m91yp/G8yBzt0HP4c79ItbvbNj4hk73MuUPOiDdv1rfd+KgBHBCeJSTR3XS3oYVA0LidK9HgaxEbbOCDAgIMbWUldv14gRF6CoCJq5EtRuoFmTA6eOwexmsfgamPKB3rq3R+bRGzjLt/KL9HH7JuOmbpjv3P0ff7l/tWQFcVbl+n2wpiW9Logfp1DfH99evSgX9ZSV/F5xxc8eO35Ih/wN/PGK7Ut36nKxpUryQBHDCOx221D9t+GaqlJ4HJytR265glx6O9ubFH0rpYVSTV6I2CVRzt+nb+AmQ9n+6nuOQy/T2ypNunf+tq2n1l4zQnjqI2b/asxIwl+gFBh3+YmYN2o5l2AZwhZl6Tmfc8I4dvyVKNU0z1KOPTsdzwnsDOEkjIjxG47QTzTrys55Y3HOE7faYQfoboTeWxHGmgoz6IWhvFt4JueDytoOPP8z+ty6XtfQWXYw7e42+340XMIhm9D9XD4FXV7q6Ja1nTSHSUhWGlkRbgrbGX47z9AIGepobwLXqs8AvQM+78+IeOAnghEdoU14i0PPfYoc2/VYWO0QPX5Udc15jvc2p43qYxaT5b24tIlHPmXFmgJ+7TV/Lbj3gmsV6kc2H10HGVzqANHGYWnSi/ufqklRHNru6Ja3X0SoMVoGhuser8UKG3G36y0q0ee8dbfosCIuXHjghXK3ZYtONGYbugetjZy6KNQiReXCtZ31TdlIJLbcSnghVp5wb4Odur+8Z7tELrnkfSo7AgR8g8WznnVc4V79JgKqfPO8JinMApYOvjopOadoDl7tNv2/4BXT8+BZt+iwIj9fJ3L2UBHDCI1hXhLUqA3rRAV2g2d5k4pgGK1FF65i8AtWtNUwl4gxlx6A013ZOUPwZcOnz+v/9z3XOeYXzBUfqoUJPCuBOHNKLu8zIkxYzCI7t1clzrfK2mz582qbPgrB4/eWotsbUNrgLWcQgPEKb0k5YKzD0Htv0vtCeEBhWH5SIlhVkgH8whCW4uiXOV5dK5ICuoWs2ywKGPaofK77PrH8tj5mt58NFDzT/nKLz9J8MG16HqtPOr91rhuJDHZ//ZhWdooeQSw7rY5bm66kXjechd1CbPgvC4/Xq2JNHvXJqggRwwmPYXRFWXqITZwZ0r992ZDP4Bug5cI0pBbGDIV8CuFbLt6xA9ekCHfbhliDVWT1wlkndN3xVRn51hm11htguMETt7fqfC2v/Doc2eEauyeJD9ZVqOsraQ38sQwdO1tXWTliB2urVwdYvnSeaJiv2Bl3gHVl4tTenwTNJ8K/rYeuHeuj0yBb9rc/RvIuYQdID1xYFGV1j/htAYAgERzuvqH3uNkoDYsivDmndHB7hWRIm6tXvnjCMWlsLxYfNC2waF7WvW4Fqbg9cm4R7dy44pwZwSqkDSqltSqktSqlNlm2PKqUOW7ZtUUpd3GD/PyilMpVSGUqpGQ22X2jZlqmUeqDB9v5KqfWW7R8ppcybKSncX1W5/rYXnaKHTT+9Bf6aDAfXNp9MM2aIruspK1FbVl4MJ490jflvVhFOTCWSu52a2OGtn8MjPEu3Hjp5eCcEcG1Kq2TPqWNQU2He1Iju0ToZtrWofe42vTgiONKc47eHNZnvCe9cyNAZQ6hTDcNo/En5N8MwFjbcoJQaClwLDAN6A98ppawZAV8GpgM5wEal1OeGYewEnrYc60Ol1KvAr4FXnPhchDux/lGedScMv0oPne78TL95Dr3c8ePqVqLuhu6TnN9OT1bQhVagWoUn6F5cs1VXwLEMws6aweLpblZGTJin3znw0wtQcbK+DrPJrKk0KqtrbYfhG6utgcoyHVg2dsKkFCJWSuleOOt7Ru525yfwbUlAsO5RlwDO6WYCHxqGUQHsV0plAmda7ss0DGMfgFLqQ2CmUmoXcB5wvWWfd4FHkQCu6yg6oG8j+un5WX3HtW7iuTUhbf4uy9J/4VBXWoFqFZ4Iu5aZX5i8IENPqO453P3KiAnz9D8XfnxOJ/VNaVpj1Az2UmmkJkboWqRHf4HsH3Vi6INr9UjF//4MYY1ShViHFc1axAC6CsPuLy2jI3tg8CXmHbu9wuO9dgjV2QGcASxXShnAa4ZhvG7ZfodS6gZgE3CPYRhFQB9gXYPH5li2ARxqtH08EAWcMAyj2s7+NpRSvwV+CxAXF0daWlpHn1ed0tJSU4/n6TrzevTJWcFA4NGvD9A35iTJEa38sDUMJvkGk7flv+w95dxVf57++hiQuYLePgH88MsBUB17E/SUa9GroIJBtVWsXb6Uim4xph03Lve/DAE2ZJ/m1LG0dh/HU65jZ3G36+FTU8kk5cfh1YvJOuKcWT2BJ2rwU1BtgK+CwBPZrP3mZ8ZtWoB/dSkAp4L6UBo6gtiCH9n91Svk9ppW9/jS0lIyD6WRDPy4LZvq3ebMw+xb7EfyqUK2fvEKo4wadhT6UODi382wym50L8pgQzPtcLfXUGs5O4CbZBjGYaVULLBCKbUb3UP2BDq4ewJ4FrjJmY2wBI6vA4wbN86YMmWKacdOS0vDzON5us68Hnn//pxTRiD/3B9MwKFKx8MI9mQNp4/fSfo4ua0e//rIeQliBzFl6vkdPpTHXIvMGtjzCkdPGyQOGWVeT9k334JfEGdedF2HevY85jp2Ere8HgcnEF+xj3gntWsKMGZske0w/LYlsK4ULl4IQ2cSHBJLsGHAwoEMDsxncIO2pKWlkewbCIdCmDTtEvNqHO+tgqy3GeWbCcCw866B6GRzjt1eFStg4xamTJ7s8Hm65WuoFZy6iMEwjMOW23zgU+BMwzDyDMOoMQyjFniD+mHSw0DDvty+lm2OthcC4Uopv0bbRRdRlpvJISOWWkO1fTWfrERtna60AtViW4V+TaX8/CR/ffOf7Z8k3ljuNj18b+awrHBP/c/Vv+9Tx512itTECG6fmlz/BSNvB/j4wdh5EBKrtyml27JvVdPycCcO6Un+ZgVvUF8TdecX4N8dIvubd+z2CouH6tNeuWjNaQGcUqq7UirU+n/gAmC7UqpXg91+BVjWGvM5cK1SKlAp1R8YCGwANgIDLStOA9ALHT43DMMAvgeusjx+HvCZs56PcD89a/PIUbHtW80XMxjKCpz6BuvxKk7quSNdaf4bsDo/iP+tvpNoVcyHvo/Q/avboeRoxw5qGJas9C5MqSA6T79JgAGH1nfeOfN2cLrHAF7+4aDtl47+5+rqH8f22u5f7ITcaGHxOul3RTHEDXWPLyt1qUS8byGDM3vg4oAflVJb0YHYl4ZhfAM8Y0kt8gswFfh/AIZh7AA+BnYC3wC3W3rqqoE7gG+BXcDHln0Bfg/cbVnwEAW85cTnI9yJYRBceojhw0Zx9wWD2jZ8CroGJehs4cJ+SoKuVAO1gQlJUXzncxbTK5/ltdqZpBSsgJdS9crCForcO0ztUHJE5yiUAK5rsJaPyu+8mssVR7bz3fHopkXe+1sSCu9fZfuA4hxzFzCAXkwWZRkydfUKVKu6VCLet5DBaXPgLKtGm6R4NgxjbjOPeRJ40s72r4CvHJzjzMbbRRdQdgyqyohLGMTtE9oxxyLY0lt3SpKoOkxJYK0X28UCuIalesYlnYdP6KPw9QOw4hHoNdphhv1mUztYk5q6y4eacK5uYRDaq/5LkLOdPkFg2WF21Z7TdGVqRD+d623/KjjzNwD41JTD6ePOqU4QMwhyf3GfLytenMxXKjEIz9QwhUh7SABXx15KAkD3HvgGQIQbzGNxInu9ZjbziyKTYNY74OMPWSsdHsfhdQT9gQYQN8xZT0O4m86cZ2vp6ctUiU2nlCgFSefC/h/qirp3Ky/Q94WZ3AMH9RUZ3CWA6xYOgT2kB04ItyEBnGkmJEUR4OdDVXWt7Rt/QQZEDQRf732baHVC1IDukDABsv4L0x+3eyyH1xF0UtPwRPsJVYV3ih4EWxbrYXczFwrYY+nhvePamYzKD2yaILr/ZPj5fb2wovdo5wZwQ2dC4V7zaqx2lFL6eXphD5z3vjML72YN4MJbVwYmPbvRkvsgS3kXCeBshgxt3vgLdkOfVNc2zskcJkS1Z8B5sPIxPW/SusqvAYfXEWQBg5dr8v4COqltZSmUmFhv1JH8nRAYxsihQxk5zE6w2P9cfbt/FfQeTWCFNYBzxhBqClzxesv7dabweOmBE8JtnDgAIT11qZQWOOxlCQiRVagWTSoDVJXr8jOjrnNdozpBs71mjVkDuKzvYdQ1dnexW2GhsgwKs2DELBNbLtyFw/cX69zRgt3OD+DydujheUc9faE9dY/g/tVw9l10K88H5avn6XUFYfG6KoWXkTlwwjMVZbd6+NTh3KTgSOmBc6ToAGBA1ABXt8SprL1mrVrJ3HOkrquY9d+2nSR/F2DIAgYv5fD9xToXrMDJCxkMQ7/G4oY2v1/SZF1eq7qSwIpj0KO3V0+PsBEeD+XF+p8XkQBOeKaiA60O4Ky9LE0m9wZHSQDnyPEsfRvp3QEc2EmI6oiPDwyYqgO4FtKJ2LAuYOgpAZw3cvj+0j1aT9U4luHcBhQfgoqSlhfI9D8Xqk7B4XQ9B87ZvYLuxEtTiXSR8Ft4lepKncOolQGcw7lJwVF6Kb1oqtASwEUlubYd7mbAebDt33rIqrUBWe52vQouPNG5bRMu4fD9RSnLSlQnB3B5lrSosS0EcP0mAQr2r9IBXEIXWhFtnStdfMirvkhJACc8T/EhwGjTClS7c5OCo6Aw09SmeY3jWbr3IMikOqDeImmqvs1a2YYA7pfm5ycJj2f3/QV0ALfzM+euRK0L4IY0v19QhF4ZmvW9HkKVHjiPJ0OowvMU7de3ER3s0QiOkkUMjhRmef38t3bp0Qtih7Z+Hlx5CRzeDAkTndsu4Z6iB+kKHM6sw5m3Q/cwtSZFTdJkOLQORa1pVRgcVh9xJyGx4NfN68ppSQAnPE9Hc8BZBUfquSPVlR1tkfc5vq9LzH9rlwHnQfZaqDzV8r77V4NRA8nnO79dwv3EWIq7d3AeXLNBUv7OlodPrazpRMCUHHDWFbhNyne5G6V0j6P0wAnhYkUHwDdQpxHpCGsyX5kHZ6vylM5dJT1w9g2YCjUVcHBNy/tmrdTpavpKxb8uqS6VSPsDuGaDpOoKXaS+tRU+EibqiiJgSgDXbPURd+OFyXwlgBOep+iAHj716eDLV6ox2Gcdoo6UBQx2JZ6tv0Bkfd/8foYBmSuh9JvMOQAAIABJREFU3zngF9A5bRPupUcfHcB3IIBrNkgqyNA9vC2lELEK6A59z9D/D+vT7jZZOVyB6468MJmvBHDC87QhhUizJICzr24FqvTA2eUfBIln6eCsOcf3wYlsGT7typSC6JQODaE2GyRZFzA4yDFod+h19PUUhY+EwNB2t8mqTXkUXS0sAcrydZJyLyGrUIVnMQydxNeMSeESwNlXlwNOeuAcGnAerHgYSo7ohKj2WBc6DDiv89ol3E/MINiX1u6HN1uiLX+H7g22M1/VYYWIsXPZWhLPlHa3qGn73Dpws7Iu2ijOgehk17bFJNIDJzzL6SK98EB64Jzn+D5dcaBbmKtb4r6sQVlzw6iZK/XrVHoyu7boFDh5tENVABwmm87boQNEOxUVPGp+WmewzvnzopWoEsAJz1KXQqRfx49lzXEmqURsFe6ToKMlccMgJM5xOpHqSjjwg/S+ifqFDMf2mn/svJ0OFzB41Py0zmDtgTvhPQGcDKEKz1KUrW/NCOB8/XUvk/TA2TqeVZ+wVtinlA7O9i7XwVrjRQo5G6CyFAbI/LcuL8ZaE3U39B1n3nHLCqE0V+cltKPZodeuKLQ3KF+vWsggPXDCs1hzwJlVlkjqodqqLNPDPVJCy4bdyeAjrtKvnQ2vN31A5krw8bPNuyW6pvBEPU/N7JJa+dYFDI5TiLS6zm9X4OunVwV7USoR6YETnqXogJ6fFRhizvEkgLN1fJ++lSS+dRxOBk+eBgMvgLSnYMQsCI2rf1DWSp37rTXZ8YV38/WDqGTzA7i8lgM40Uh4vFcNoUoPnPAsZqUQsZIAzpakEGmi2cngM/4Pqsth5eP120oL4OhWmf8m6sV0LJWIXXk79PtXSFzL+wotMqn+Pc4LSAAnPItTAjhZxFBHUog00exk8OhkmPg72PI+5KTrbdaUEckSwAmLmMF6/m7VafOOmbdDz39TyrxjeruoZJ0LrgMrgt2JBHDCc9RU6Rw+pgZwkdID11DhPv2N3oQkn96ixWSl596nr9nX90FtrR4+DYqEXqNd02DhfqJTAMO8lai1tXpRhIMEvsKBKEv+Ny/phZMATniO4hxdNsbsHriqU60rTN4VHM+S+W92NDsZPDAUpj8Oh9Nh6wc6tciAqeDj2/kNFe6pLpXIHnOOV5qr37dkqkPbRA/UtxLACdHJrCtQzQ7gQAraWxVmyQrU9hhxtV608NX9UJon89+EragBoHzMW8hQfFjfmlCQvkuJ6Kd/D4VOyMnnAhLACc/hzABOhlGhvETPD5EeuLbz8YGLnta9IiABnLDlFwgR/fWwpxlKcvStCQXpuxS/QAhPgMJMV7fEFJJGRLhcenZR65JNFh0AH3/HtSfbQwK4etYUIjIs0ypNXrd9xsKE3+kPaTNfo8I7xAw2bwjV2gPXQwI4Rxx+rkQlSwAnhBkc5tiy59gencfHzLlFdQGcDKHWr0CVAK4lDl+3F/7F1U0T7iomBfZ+qxdj+fp37FjFOeDfvb4coLDR7OdKVDIcXAeG4fEreGUIVbhUqwsub/0IMr7SiVPNJD1w9QqtSXxlDlxLpFC4aEmT6h2xQ6G2GvJ3dvzgJTl6+NTDAxBnafbvMypZl7krzXNdA00iPXDCpaw5tqqqax0XXM5eA5/fAf3OgelPmNuAbmF6UqsEcLoHLrQ3BAS7uiVur1WvW9Fl2e0BspZVy/ov9BrVsRMUH5bh02Y0+/dZl0okE0J7uqaBJpEATrhUiwWXC7Pgw+t1PcFr3mtaNLyjfHz1MIQEcJYVqDJ82hpSKFw0x14PUGpiMvQcAXu/g0n/r2MnKDkMcfaL2IsW/j6tAdyxvdBvkmsaaBIJ4ITLpSZGNPkATM8u4ueMLOZsv5luygdmf+y8+R5STks7vg8GX+LqVngMe69bIaCZHqDkabDmJb3iu711cqsr9PCfpBBplsO/zx59wK+bVyxkkDlwwu2kZxdx45s/MOLH2/EpPsTuKa85d16WBHC6tMypY9IDJ4QJHFbvSJ6u58HtX9X+g5cc0bcyhNo+Pj56oZYXJPOVHjjhdtbtK2Se8QXjfXezoOoOBpb1Z7AzTxgcVZ9jrguwu7y+UFagCmEmuz1A8WdCYA/YuwKGXNa+A5dYk/hKANduUQMgf5erW9FhEsAJtzMhKYoTaVnsqe3LNz6TmOvsCeLBkboMUhfgcHm95IATwvl8/SFpMmSubH8ai7occH3NbVtXEpWssxrUVIOv54ZBMoQq3E5qYgRnhxfhG5vSfF44s1iHUA3DuedxAw6X1xdmAUpnixdCOE/yNJ0GpL1VGaQKQ8dFD9RD2SeyXd2SDpEATrifmiq6nTzIgMGjO2eSeHAU1FTq3EBezjq52ldhO7n6eBaE9QX/bq5toBDeLnmavs38rn2PL87RC7oCupvXpq6mYSoRDyYBnHA/Rdn621H0wM45XxdK5mt3cnXVaTjwky71I4RwrrC+EDNEz4Nrj+LDMnzaUV4SwHnu4K/wXoV79W2UCwK4iH6dc04XajK5eu3LeljmV6+6rlFCdCUDp8H616CiFAJD2vbYksM6CBTtFxypezE9PICTHjjhfo5ZArjo5M45X1euh3oyF354DgZfCv3PcXVrhOgakqfpaRsHfmj7Y4tzJIAzgxcUtZcATrifwr0QHN15hZqDI/VtFxhCbWLlE/qD5AKTS5QJIRxLmKiL0bd1GLWiFMpPSA44M0QN9PhccBLACfdzbG/nzX8Dt5sD16QItrMc+Rm2LIYJt0kBeyE6k18g9D8XMle0bfV7XQ446YHrsKgB+npWlrm6Je0mAZxwP8f21k8y7QyBPcDHzy0COGuetmeXZzD7zXXOC+IMA775ow5ez73XOecQQjg2cBqcOMjir1a2/u+82JJCRHrgOq5uIYPn9sJJACfcy+kiXdKpM3vglHKbcloO87SZbdfncHANnPcQdAtzzjmEEHUa96xvCzoTgKy1/2n9lzXpgTOPF6xElVWowr0cs/wxRad07nndJIBzWAS75IgOtNqY+8lu2ayqclj+MMQOg7E3mPwMhBCN2auAsq4gmODaXvyPzxqoVfgu/wJCi3Vy2f6TYcaTTQ9UnAMo6NG705+D17FOGynMAiJd2pT2kgBOuJfOTiFiFRzlFqtQrXnabIIuw4DXp8Kwy+Gip1t9LIdlsza+qT8kbvgMfHyd+GyEEGC/Z31CUhRp36fya59ljPbJoiYvCCr7QdUp2PQOTHusaZmn4sMQEqdLcomOCQjW+fQKMyHyDFe3pl0kgBPu5dhePR8tIrFzzxscCQUZnXtOB5rkaTt5FEpzIfunNh3H3odGamKErgHYazQkTTG13UII++z1rKcmRqDmLeSjHdcxeNBgRqUM0NM5tn8CS26C3K3QJ9X2QCU5UkLLTNGWVCKe2QEnAZxwM4V7dT3Ozv6G6SZDqHbl79K3eTuh8pT+5tgKdodjqysgZxOc+RsnNlgI0ZDdnnVg7IBejB1wqe3OCWfp24PrmgZwxYchbmgntLiLiEqGbf+GZM+sg92qAE4pFQQkGIbhHl0Uwnsdy+zcBQxW1iHU2lrwcbO1Pdai10YNHN0KiRNb9TC7HxoH10NNBSRMcGKDhRCNNelZd6RHL10RJnsNTLy9frth6DlwAy9wWhu7nKhkKC/Gv6rE1S1plxY/qZRSlwFbgG8sP49WSn3emoMrpQ4opbYppbYopTZZtkUqpVYopfZabiMs25VS6kWlVKZS6hel1NgGx5ln2X+vUmpeg+2pluNnWh6r2vb0hVuprYHj+zo3hYhVcJQOkCqKO//cLcnfqZN+Ahze1KaHpiZGcPvU5PoPjoNr9G1C64JAIYQLJEzUPXANc8T9//buPD7K8t77+OeXhARkDSEsEgiyFFxRghiXWrR1b9W29qjHtS21i572PD5d9GlPF3vsOd1Ol6NdLFWr0la7uNTjhpUcrRqXKCIqSEDCIhAIAUxYkjDX88d9D5mEyWSSzMw998z3/XrNazLXveSXi0nmx7XuaYaOPepCTSX/s2bInncDDqR/kmlq+DYwD9gB4JxbChzWh+9xmnPuWOfcXP/1DcDfnXMzgL/7rwHOAWb4j2uAX4KX8AHfAk7w4/hWNOnzz/lMzHVn9yEuyTY71nmtQ0G1wEFWTGQ4SOMKmDgHRk6GjXUDu1fD894M36FjUhObiKTe5BO95ZRil7jQGnCpVzYNgFVr16V/4fQ0SCaBa3fOdW+WGEiH8QXA7/yvfwdcGFN+l/PUAqPMbAJwFrDYObfdOdcMLAbO9o+NcM7VOucccFfMvSSMon+sMj0DFbJ3Oy3nvC7UsYdDRRVsGEACF4nA+lp1n4pku2gLecNznWUH1oCblPl4clTdzuG0uUL2bV2V3oXT0ySZMXBvmNk/A4VmNgP4IvBcL9dEOeAJM3PAr51ztwHjnHOb/OObgXH+1xOB9THXbvDLEpVviFN+EDO7Bq9Vj3HjxlFTU5Nk+L1raWlJ6f3CbiD1UbH+UaYDz65spP2d/t2jv4a9t5a5wOsv1NC0enfK7jvQ90fJ3kZObGvh7eYCCiKjmL5zHc8+8QDtxaP6fK+hLWs5fu9O3tpdypYA3rP6XUkN1WNXOVkfznHSoJFsf/F+Vrznzcg/dOMS3gc890YDbfXvDej2OVln/fDw6jbWR07g8sIn+Uf70fzhyULem1YcdFhJSyaB+xfg68A+4PfA48C/J3n/U5xzG81sLLDYzFbEHnTOOT+5Sys/cbwNYO7cuW7+/Pkpu3dNTQ2pvF/YDag+/vYADB7FyWec702nz6QdU6EOjp52KBw3P2W3HfD74+0noBbed4rfuLz6Dk6uLIGZ/bjni78B4PCzPsXhpVP6H1M/6XclNVSPXeVsfWx5P+O3LGd89GdbvARWD+KkMy4c8ESrnK2zXnRf2Hz4Yc18euFnmeK28LNBt7B29gc4fO78oMNMWq/vAufcbufc151zx/uPbzjn9iZzc+fcRv+5EbgfbwzbFr/7E/+50T99IxDbNlzhlyUqr4hTLmHV5M9ADWIuSpZtaH/AVn8JkbGzYMIxYIX9Hwe37nkYPgFGZXiNPRHpu8knQvNa2OV3WO3a6M1QzbZZ8iERb5/pqspSfrvgAzxQcQMFw8dy+JLPQHND0KEmLZlZqIvNbFTM61IzezyJ64aa2fDo18CZwHLgISA6k/Qq4EH/64eAK/3ZqNXATr+r9XHgTP/7lvr3edw/tsvMqv3Zp1fG3EvCaNuqYMa/AQw6BIoGZ18C17gCho2HIaXeNlpjj/DWcesr57wJDJNPDCZBFpG+iS4XtO5573nnRo1/G4Ce9pmuqixl/vvGUnLVX71JdIs+AXt2BBxtcpJJ5cc45w78NP5EgrFJXDcO+IeZvQa8CPyPc+4x4D+BM8xsFfAh/zXAI8AaoB74DfAF//ttB74LvOQ/bvLL8M9Z6F+zGng0ibgkG+3d5e02MCaAJUQgqza076LxTa/1LaqiCt59xZuQ0Bc71sF770LlSamNT0TSY/xsb/mgaAK3a4NmoA5AdGHzQqPrPtNR5TPh4nu8pazuvRw62oIJtA+SGQMXMbPJzrl1AGZWSRKzUJ1za4DZccqbgA/GKXfAtd3L/WO3A7fHKX8ZOKq3WCQEgpyBGnXI6OxaRiQSgW1vQ9XVnWUTq6DuTti+um/LrUQ/BDQDVSQcCougYq7Xch7ZD7ve1RpwA9DTbhhdHHYqXHAL3P9ZeOFXcPIXMx9oHyTTAvd1vJa0u83sHuBp4Mb0hiV5J5rABbEGXFS2tcDtaPA2ti6PaYGb6C+n2NdxcOueh5KRXhesiIRD5UmwZbn39zHSoRa4ATpoYfN4Zl8Ck06AV+/uupByFkpmEsNjwBzgXuCPQJVzrtcxcCJ9sm0VWAGMnhpcDGlO4Ooamrl1SX3yaw1Ft9Aae3hnWflMKB4WfxzctlVw98c6906N1fA8TD4BCgr7HriIBGPyiYCDN+73XmsMXGYcd7nX+9Gf8cYZ1GMCZ2az/Oc5wGTgXf8xOXabK5GouoZmHl7d1r/FEJtWwajJUFSS+sCSNXSs103Rlrp14KLizYDqVeOb3nP5zM6ygkI49DjYWNc1IXQOHvkyrP473Hcl7GvpvKa1CbatVPepSNhUzIWCIm/DdVAXaqYc+VFvYturdwcdSUKJWuCu959/HOfxozTHJSETTVD+sqq9fytab6sPdvwbwOEf9ros30z9ZOaeZkAl1LgCRlTA4JFdyyfOIbL5dT658OkDCWH9M/fCmho4+p+87pb/ub6z+f/A+DdNYBAJleKhMGF25xATdaFmRslwOOJCWP5XaGsNOpoe9ZjAOeeuMbMC4BvOudO6PU7PYIwSAtEExdGHBCUqEvHXgHtf2uJLSuXJXhJZd0fKb93rDKh4tr7VdQZq1MS5FETambZ/LREH1rGX8udugvLD4cJfwPwbYdm98Mpd3vnrnofCYq/lTkTCJbqt1qBDvOWEJDOOuxza3oO3/hZ0JD1KOAbOORcBbslQLBJi0QSlgD4kKFG7NkLHnuCWEIky82Z8rn8BtryZ0ltHZ0Bdf+ZMFi2oTjyIFrxZZ1vf7jqBIWpiFQBzi1ZTaHDNoEcZuXcjb8/5Brc+3UDd5E/B1NPg0a/C5uVeAjexCgYNTunPJCIZEE3gRlZoDcdMqjwJSg+DV+8JOpIeJTML9e9m9nF/sVyRuKIJysdmDOo1QTloMH/TKu856C5UgNmXeq1Vr/wu5bdOagZU1PZ3vEUl480aHTkRhk/gc9N38G8fKOWLxX+jefJZnP9IodelevtLvDbvhzB4lDcebtNrnR8CIhIu0bGr6j7NLDM47jJY+4z39zgLJZPAfRb4E7DPzHaZ2XtmtivNcUkIVVWW8uFpxV0TlM2vw96dB152H8z/2oq34Yl/g8ISGHdkAFF3qmto5tYXm9leeTa89gdo3xNcMLFbaMUzsYqyHa9zdesdFLr9/G38tV3G2P1jk8FFv4Xmd7zlB5TAiYTT0DEw7XRviIdk1uxLAYOlvw86kriSWUZkuHOuwDlX7Jwb4b8ekYngJOR2rINfnwq/PAXW1QJdB/OP6Whk8oMXQdNquPQP3kK6AYlNLL+06lgv6XzjgcDiodFfQmTMzPjHJ1Z5K4a/fh+c9C8ceeQxB4+xm3IKfOjbMGS0t4SIiITTFffDB74SdBT5Z2SFlzwv/b03rCXLJFpGZIaZPWhmy83s92am9lvpmzcfAhfxHnecAzXfp3rKSIqLCphWsIl7i29iRMd274/T9IM258io2MTyuY6ZNA+p9HY8SLeWrfDu0oPLG9/0llUpGRb/On8cHMMPhfdf3/MYu5O/BF+pP3gmq4hkpT6vFynpddzl3jZm7/xv0JEcJFEL3O3Aw8DHgVeB/85IRJI73noIxh8NX3gejroIar5H1ZIr+NvZe/jb0O8xdnCEwk8+3Llpc4C6zhItpOWoy2F9bfxFcVOp5nuw8IOweknX8q0rEu+aUDEXxh0N5/3IW2qABGPstHivSCj0a71ISa+Z53rjibNwMkOiBG64c+43zrmVzrkfAlMyFJPkgl3verM5D78ABo+Aj/8GPvpr2Pw6MxZ/kkMGlzBowWNw6LFBRwocPEt00vxPe5MZ0t0KF90i576rYOtKr2x/u7erQrwZqFHFQ+Hz/4BZ56U3PhHJmH6tFynpNWgwHPNP8NbDsCe7EupECdxgMzvOzOb4Oy8M6fZapGdvPew9H3FBZ9nsS+Bzz8AJn4dPPdp1h4Es0KUFa2gZHH4+Ha/+nl89ubzzf8KRiDdmL1W7NTQ3eIOTi0pg0SegdZs3ti3S3mULLXWriOS+fq0XKel31Me9VQHWPht0JF0UJTi2CfivmNebY147QIv55rNX7oI9O+DkL8Y//tZDXgtSebfFeUdPhXP+M/3x9aKuoZnaNU1UTy3rcVmPlRUfZ+byP9NR80OefrqE6ZO2MLLpVW+Cw7xr4NwfDiyI/R2wcwMc/Qk44ya48zz442Uw91PecT+Bi3artHVEKC4qSG4dOREJnWhPQG9/myTDoiskbF3h7diTJXpM4Jxzp2UyEAmRjn2w+JteIvO+sw5uSWvZCg3Pwvu/nFSilGnJJkRP7p7BoMgErivyZqNub57qba+ysQ7eeWbggezaAG4/lFZ6Y9ou/CX8+ZPeuDsrOLAzRbxulWypSxFJrarKUv1+Z5uS4d62htFhLlkimXXgRLp6+zFvLIBzsOR7Bx9f8TC4CG+Wzs/KAbnJjjOpnjaGL7iv8Kn2rzJv/0LeufgpOP/ncPj53v/EYta365fmBu95VKX3fNTH4PR/g307oXQKDBrixaFuFRGRYJXP9P7uZ5FEXagi8S39AwyfAMf+MzzzY28ZjNjJCG89BKOnsqR5LG0dO7Ku5SiaELV3RBImRFWVpdy84GPUrmni2tgWxEnHAw42vDyw5U92eAncXSvgyMJm7/7v/79eYjh0TJc41K0iIhKg8llez1Jkf9bM7FcCJ73q0g1a1g6rnoCT/sVbY+yl38JT/w6X/xmAovb34J2n4cTrqJ42huIl9b0mSpnWl4QobnfGxLmAwYaXBpTAbVq7gnJXwHef2Unh87WdXblnfje5OEREJOXiDv0pnwkde70F6kcfFmyAvh4TODM7C28pkT93K78I2OmcW5zu4CR43ceLLT5xOZPcfq/1bfBIOOVf4clvQ8PzUHkiY7a96C2LccT5VE3M3pajASVEg0d4a7Stf3FAMby3uZ52ymh3hUSyqIVSRCRf9ThGOroqwNaVWZPAJRoD900g3tLDNcBNaYlGsk738WJD3rgXDp3TOXFh3jUwbBw89V1wjjHbnoeRk7xz6OMG7mEy6XivCzUS6fctDnWNbGSsxraJiGSJHsdI+5PKsmkcXKIErsQ5t7V7oXNuGzA0fSFJNokdQH900TrGtK7yWt+iiofCqV/xxga8+SCjt7/qDfI3Cy7oTKiY50022PZ2v28xbPcGps88+uDtr0REJBA9ThobMsob+51FM1ETjYEbYWZFzrmO2EIzGwQMSW9YEpTuff+x48Uu2vokrCz2FjWMNecqePbn8MDnKXBe92nOmzTPe97wIoxNsGNCT9paoXUr5RUzuPbU6amNTURE+iXhGOnymbA1zdsr9kGiFri/Ar8xswOtbWY2DPiVf0xyTE/78FVVlnLtqZWMW/sQvO9sOGR01wuLimH+DdC+m33FpV7rVK4rmw5DSvs/Dm7HOu+5dErKQhIRkYHrcehP+SzY+vaAhs6kUqIE7hvAFqDBzOrMrA54B9jqH5Mck3B9tFWLYfc2OPay+BcfczFUzGPThLOgIA+WFzSDiuO9maj90bzWe1YCJyISDuUzob3VW4Q9CyTqQn3EOXemmX0HiPbx1Dvn9mQgLglAwvXRli6CoeU9L5tRWAQLFrO2poYpGYk2C1TM85ZU2bPDGx/RF90X8RURkexW7g+X2boSRk0ONhYSJ3DlAH7C9npmwpEg9dj339oEbz/uzTgtHBRskNlk0vHe88aXYfqH+nbtjgYYdEiXBXtFRCSLHUjgVsCMM4KNhcQJ3Egz+1hPB51zGgeXg+Kuj7b8zxBp7zr7VGBilbdn6fqX+p7ANa/1Wt9yfbauiEiuOGQ0DB2bNUuJJEzggA8D8T5hHJrIkB+cg5fvgAmzYfxRB4qzcZP6jCsZ7i3ou6EfExmaGzT+TUQkbMpnZs1SIokSuAbn3KcyFolkp3XPe9OmP/LzA0U9rlSdjyqOh+V/8WYlJTt5wzmvC/Ww96c3NhERSa3yWbDsXu/veMA9KIk+cdS3I95epyUj4eiLDhQlnK2abybNg327+takvrsJ2lo0gUFEJGzKZ3p/89/bFHQkCRO4K7oXmNkYMw3aCau6hmZuXVJ/YH23XrVshTcfhNmXeDsu+HpcqTofVcQs6Jus6AxUdaGKiIRLdCJDY/AL+ibqQh1mZjXAduC7wN3AGKDAzK50zj2WgfgkReJ2e04eBe++AhOOhYLCgy9aeo83eWFu1570hCtV55uyaTBkNNtW/IN7d52SXH3sWOs9l6oFTkQkVGKXEulpWa0MSZTA3QL8P7zJDE8B5zjnas1sFvAHQAlciMTr9qwqegd+czqc/CU446auF0Qi3uSFylPibhUVd7ZqPjJjR9mx7Hz7WX68fGVyYwKji/iqC1VEJFyGjoEho7NiJmqiLtQi59wTzrk/AZudc7UAzrngo5Y+i9vtGW0CfvZnsOKRrhes/rs30P54zWPpzYqiWUyzdxnuWpIbE9jcAIeMgZJhmQlQRERSw8zfUiv4maiJErjYzb66777g0hCLpFG02/P6M2d2thA11UNBEYw/Bh74HGx/p/OCl37r7bww6yPBBR0SpbNOAeDUwuXJjQnc0aDuUxGRsCqf6bXAuWBToUQJ3Gwz22Vm7wHH+F9HXx+dofgkhQ7aoLepHkoPg4vv9l7/6Spo3ws71sOqx+G4K7yN6iWhmXNOo23oofxs0C3879FPUDU+0cgEvBY4dZ+KiITT2MNh7w5oaQw0jB4TOOdcoXNuhHNuuHOuyP86+lr7KeWCptVQNt2bDXnhr2DTa/DYDVB3p/c/i6qrAw4wJIoPofi65yiouopxb94BtxwPb9wf/39nkf2wc71moIqIhFX5TO854HFwSa48KjknEoHtq71ZlACzzvUmM9TdAc/fCjPOVDdfXwwphY/8FBY86Q1y/dPVsOgiCjt2dz1v10aIdKhuRUTCKnYmaoCUwOWrXRuhY6/XAhd1+jeh8mTo2APHfzq42MKsYi58pgbO+g+of5LJ6/7S9Xh0DTh1oYqIhNOwcTB4ZOAtcL0M1pGc1bTKe45N4AqL4J/uhjVLYPoZwcSVCwqL4MQvwIaXqHjrYW9B5GHl3rHoEiJqgRMRCacDM1HVhSpBaFrtPY+Z0bV8aJm3bVay+3pKz077fxRE2uAf/9VZtqNBuQytAAAbxElEQVQBrABGTgouLhERGZjoTNQA6VM6XzXVQ/EwrylY+iTpLcnGzGDz+NO8JVl2bvTKmhtgRAUUah6QiEholc/y9rVu3RZYCOpCzVdN9d4EBm1t2ydxtyRLsOtCQ+XFTGh8Gp7+oTfJoXmtuk9FRMLufWfDyAoYNCSwENQCl6+a6ruOf5OkxNuSLJG9Q8ZB1VXw6t3eQslaxFdEJHQO6nkpmwZHXADFQwOLSQlcPurYBzvWKYHrh7hbkvXm/V/2drx48tvQsgVGTUl3mCIikiLRnpcfP7GSyxbW9j58JkPUhZqPmteCiyiB64folmS1a5qonlqWeNP6qBET4PgF8Pwt3mu1wImIhEa8npek/vanmRK4fNRU7z1HF/GVPqmqLO37L+8p13s7XLS1aBcGEZEQifa8tHdEku95yQAlcPkomsCNnkZdQ3PfWpOkf4aWwYnXwTM/gtFKnEVEwqJfPS8ZkPYEzswKgZeBjc65D5vZncAHgJ3+KVc755aamQE/A84Fdvvlr/j3uAr4hn/+vzvnfueXVwF3AkOAR4AvORdvA0rpoqkehpZT1+j6NKNSBugDX4PZF3vJnIiIhEa/el7SLBOTGL4EvNWt7CvOuWP9x1K/7Bxghv+4BvglgJmNBr4FnADMA75lZtFa/CXwmZjrzk7nD5Iz/E3s+zqjUvqmrqGZh1e3dQ54LSiA0VODDUpERHJCWhM4M6sAzgMWJnH6BcBdzlMLjDKzCcBZwGLn3HbnXDOwGDjbPzbCOVfrt7rdBVyYnp8kx2xbBWXT+zejUpISnbX0l1XtWTVrSUREckO6W+B+CnwViHQrv9nMlpnZT8ysxC+bCKyPOWeDX5aofEOccklk705obYSy6Qf69a8/c6a6T1Ms2rrpUOumiIikXtrGwJnZh4FG51ydmc2POXQjsBkoBm4DvgbclK44/FiuweuWZdy4cdTU1KTs3i0tLSm9X7oN37WKKmD5pr1s8+M+0uC9dzZQ887A7x+2+kiXkh37KTLocI5CM0p2NFBTs6H3C3OY3hupoXrsSvXRd6qzrsJaH+mcxHAycL6ZnQsMBkaY2T3Oucv94/vM7A7gy/7rjUDsDt8VftlGYH638hq/vCLO+Qdxzt2Glywyd+5cN3/+/Hin9UtNTQ2pvF+q9Di7dNlWeAWO+sCFMHZWyr9vttZHps0HjpvTzB+efIlLP3S8WjfReyNVVI9dqT76TnXWVVjrI21dqM65G51zFc65KcAlwFPOucv9sWv4s04vBJb7lzwEXGmeamCnc24T8DhwppmV+pMXzgQe94/tMrNq/15XAg+m6+cJk4SrRjfVAwajDwssvnxRVVnKh6cVK3kTEZGUC2IduEVmVg4YsBT4nF/+CN4SIvV4y4h8EsA5t93Mvgu85J93k3Nuu//1F+hcRuRR/5H3Eq4a3VQPoyZDUUnim4iIiEjWykgC55yrwev2xDl3eg/nOODaHo7dDtwep/xl4KhUxZkrEq4arU3sRUREQk87MeSgHleNds5bA27SCcEGKCIiIgOiBC5HxV01uqUR2t5TC5yIiEjIZWInBskW2sReREQkJyiByyfRBG7MjGDjEBERkQFRApdPmlZBYQmMqOj9XBEREclaSuDySdNqr/u0QP/sIiIiYaZP8nzSVK/xbyIiIjlACVy+iOyH7e/AaCVwIiIiYacELl+0boNIO4ya1Pu5IiIiktWUwOWL1kbveWh5sHGIiIjIgCmByxetW73noWODjUNEREQGTAlcvmiJJnBqgRMREQk7JXD5ItoCN0wJnIiISNgpgcsXrY1QWAwlI4KORERERAZICVy+aN3mjX8zCzoSERGRUKtraObWJfXUNTQHFkNRYN9ZMqt1KwwdE3QUIiIioVbX0MxlC2tp64hQXFTAogXVVFWWZjwOtcDli5ZGTWAQEREZoNo1TbR1RIg4aO+IULumKZA4lMDli9ZtMExLiIiIiAxE9dQyiosKKDQYVFRA9dSyQOJQF2o+cM6bxKAuVBERkQGpqixl0YJqatc0UT21LJDuU1AClx/27YL9bVrEV0REJAWqKksDS9yi1IWaD7SIr4iISE5RApcPtIiviIhITlEClw+0kb2IiEhOUQKXD1rVhSoiIpJLlMDlg+gYuEM0C1VERCQXKIHLB61bYchoKNSkYxERkVygBC4ftG7VIr4iIiI5RAlcPmjdqvFvIiIiOUQJXD5QAiciIpJTlMDlgxYlcCIiIrlECVyu69gH+3ZqEV8REZEcogQu12kNOBERkZyjBC7XKYETERHJOUrgct2Bjey1jIiIiEiuUAKX6w60wGkXBhERkVyhBC7XRTey10K+IiIiOUMJXK5r3QaDDoHioUFHIiIiIimiBC7XaRFfERGRnKMELte1NCqBExERyTFK4HJd6zaNfxMREckxSuByXWujZqCKiIjkGCVwuSwS8Vrg1IUqIiKSU5TA5bI9zeD2axFfERGRHKMELpdpEV8REZGcpAQul2kRXxERkZykBC6XaSN7ERGRnKQELmTqGpq5dUk9dQ3NvZ/cus171hg4ERGRnFIUdACSvLqGZi5bWEtbR4TiogIWLaimqrK05wtaGsEKYUiCc0RERCR00t4CZ2aFZvaqmT3svz7MzF4ws3ozu9fMiv3yEv91vX98Ssw9bvTLV5rZWTHlZ/tl9WZ2Q7p/lqDVrmmirSNCxEF7R4TaNU2JL2jd6k1gKFBDq4iISC7JxCf7l4C3Yl5/H/iJc2460Ax82i//NNDsl//EPw8zOwK4BDgSOBv4hZ8UFgK3AucARwCX+ufmrOqpZRQXFVBoMKiogOqpZd6Btf+AFf9z8AXaB1VERCQnpTWBM7MK4Dxgof/agNOBP/un/A640P/6Av81/vEP+udfAPzRObfPOfcOUA/M8x/1zrk1zrk24I/+uTmrqrKURQuquf7MmV27Txd/C+7/HLTt7npBtAVOREREckq6W+B+CnwViPivy4AdzrkO//UGYKL/9URgPYB/fKd//oHybtf0VJ7TqipLufa06Z3JW2Q/bHkD9u2CFQ93PbmlURMYREREclDaJjGY2YeBRudcnZnNT9f3STKWa4BrAMaNG0dNTU3K7t3S0pLS+/XVIa3rmdexB4DtS25h2fbOhO39u7bwbvNeVmcwvqDrI9uoPjqpLlJD9diV6qPvVGddhbU+0jkL9WTgfDM7FxgMjAB+BowysyK/la0C2OifvxGYBGwwsyJgJNAUUx4Ve01P5V04524DbgOYO3eumz9//oB/uKiamhpSeb8+W/YneAk4/COMfuth5h87DUZNgrZWqNnLpFlzmHRK5uILvD6yjOqjk+oiNVSPXak++k511lVY6yNtXajOuRudcxXOuSl4kxCecs5dBiwBLvJPuwp40P/6If81/vGnnHPOL7/En6V6GDADeBEvbZnhz2ot9r/HQ+n6ebLW5mVQWAwf+g7g4LU/euVaxFdERCRnBbG+xNeA682sHm+M22/98t8CZX759cANAM65N4D7gDeBx4BrnXP7/Ra864DH8Wa53uefm182L4Oxh0PZNJjyfli6CJyDlmgCpzFwIiIiuSYjC/k652qAGv/rNXgzSLufsxf4RA/X3wzcHKf8EeCRFIYaLs7BpmUw6zzv9bGXwQOfg3W1sMffqUGzUEVERHKOVngNs13vwp7tMGG29/qI86F4GCy9p7MLVRvZi4iI5BxtpRVmm5d5z+OP9p6Lh8KRF8IbD3R2nR6iFjgREZFcoxa4MNu0DDAYd1Rn2bGXQVsLvHoPlIyAQYMDC09ERETSQwlcmG1e5k1eKBnWWTb5RCg9DFobNQNVREQkRymBC7PNyzq7T6PMvFY4UAInIiKSo5TAhdWeZtixDsYfc/Cx2ZcABsOUwImIiOQiTWIIq83LvecJcRK4UZPg9K/D2CMzG5OIiIhkhBK4sDowAzVOAgdw6lcyF4uIiIhklLpQw2rTMhg2Xuu8iYiI5CElcGG1+fX43aciIiKS85TAhVH7Xti64kD3aV1DM7cuqaeuoTngwERERCQTNAYujBrfBLcfxh9NXUMzly2spa0jQnFRAYsWVFNVWRp0hCIiIpJGaoELo82ve88TjqF2TRNtHREiDto7ItSuaQo2NhEREUk7JXBhtHmZt03WqClUTy2juKiAQoNBRQVUTy0LOjoRERFJM3WhhtGmZd7+pwUFVFWWsmhBNbVrmqieWqbuUxERkTygBC5sIvthyxsw54oDRVWVpUrcRERE8oi6UMNm+xpob+15AV8RERHJeUrgsljc5UE2veY9d9/EXkRERPKGulCzVNzlQSaNhGX3QtFgKJ8VdIgiIiISELXAZam4y4PUfA9WPQEf+g4UFQcdooiIiARECVyW6r48yDk8B0//EI67Ak74bNDhiYiISIDUhZqlYpcHOW3EJqY++kmYVA3n/RjMgg5PREREAqQELotVVZZSVdYOt10Ah5TBxXdDUUnQYYmIiEjA1IUakKQ2oN+7E+69AnY3waW/h2FjMxegiIiIZC21wAWgpw3o6xqaWbqyng9aHVO2PAlraiDSDhfdDhNmBx22iIiIZAklcAGIN8O0oH03O+++gqt4lSKLsG9YBSUnfBaO+jhMnBN0yCIiIpJFlMAFIDrDtL0jcmAD+s0v/IlzrI47Os7i/sipnDXnLK49fUbQoYqIiEgWUgIXgHgb0G95cSX73CB+sP+fcUUlfGvamKDDFBERkSylBC4g3TegH7fzNVrGzea6w486kNSJiIiIxKMELhu074V3lzKs+vNce9r0oKMRERGRLKdlRLLBpqXebNNJJwQdiYiIiISAErhssP4F71kJnIiIiCRBCVw2WPcCjJ4Kw8qDjkRERERCQAlc0JzzWuDU+iYiIiJJUgIXtO1rYPc2mDQv6EhEREQkJJTABW39i97zpOpg4xAREZHQUAIXtPUvQMkIKJ8VdCQiIiISEkrggrb+Bag4Hgr0TyEiIiLJUdYQpD07oPEtTWAQERGRPlECF6SNLwMOJiuBExERkeQpgQvS+hfBCmBiVdCRiIiISIgogQvS+hdg3JFQMjzoSERERCRElMAFZX8HbHhZ499ERESkz5TAZUBdQzO3LqmnrqG5s7DxTWhrUQInIiIifVYUdAC5rq6hmcsW1tLWEaG4qIBFC6qpqizVBvYiIiLSb2qBS7PaNU20dUSIOGjviFC7psk7sP5FGDYeRk0ONkAREREJHSVwaVY9tYziogIKDQYVFVA9tcw7sP4Fb/9Ts2ADFBERkdBRF2qaVVWWsmhBNbVrmqieWuZ1n+7aBDsaYN41QYcnIiIiIZS2FjgzG2xmL5rZa2b2hpl9xy+/08zeMbOl/uNYv9zM7OdmVm9my8xsTsy9rjKzVf7jqpjyKjN73b/m52bZ2ZxVVVnKtadN95I3gHXPec+VJwUXlIiIiIRWOlvg9gGnO+dazGwQ8A8ze9Q/9hXn3J+7nX8OMMN/nAD8EjjBzEYD3wLmAg6oM7OHnHPN/jmfAV4AHgHOBh4l2zU8B8XDYPwxQUciIiIiIZS2FjjnafFfDvIfLsElFwB3+dfVAqPMbAJwFrDYObfdT9oWA2f7x0Y452qdcw64C7gwXT9PSjU8741/K1QPtoiIiPRdWicxmFmhmS0FGvGSMH/tDG72u0l/YmYlftlEYH3M5Rv8skTlG+KUZ7fd26HxDZis7lMRERHpn7Q2ATnn9gPHmtko4H4zOwq4EdgMFAO3AV8DbkpnHGZ2DXANwLhx46ipqUnZvVtaWvp0v7JtL3I08GrzEHamMI5s0df6yHWqj06qi9RQPXal+ug71VlXYa2PjPThOed2mNkS4Gzn3I/84n1mdgfwZf/1RmBSzGUVftlGYH638hq/vCLO+fG+/214ySJz58518+fPj3dav9TU1NCn+z3xJBQWc9x5n4ZBQ1IWR7boc33kONVHJ9VFaqgeu1J99J3qrKuw1kc6Z6GW+y1vmNkQ4AxghT92DX/G6IXAcv+Sh4Ar/dmo1cBO59wm4HHgTDMrNbNS4Ezgcf/YLjOr9u91JfBgun6elGl4Hg6dk5PJm4iIiGRGOlvgJgC/M7NCvETxPufcw2b2lJmVAwYsBT7nn/8IcC5QD+wGPgngnNtuZt8FXvLPu8k5t93/+gvAncAQvNmn2T0Dta0VNi2Fk74YdCQiIiISYmlL4Jxzy4Dj4pSf3sP5Dri2h2O3A7fHKX8ZOGpgkWbQhpcg0qH130RERGRAtJVWJjU8D1bgLSEiIiIi0k9K4DKp4VkYdxQMHhl0JCIiIhJiSuAypaMNNrys7lMREREZMCVwmbJpKXTsUQInIiIiA6YELlMa/A3sJ58YbBwiIiISekrgMqXhOSibAcPGBh2JiIiIhJwSuEyIRGB9LVSq9U1EREQGTglcJjS+CXt3QuXJQUciIiIiOUAJXCZo/JuIiIikkBK4FKtraObWJfXUNTR3Fq57DkZUwKjJwQUmIiIiOSOde6HmnbqGZi5bWEtbR4TiogIWLaimausD8OZDcOylYBZ0iCIiIpIDlMClUO2aJto6IkQc7O/ooGDxN2DDPTD9Q3DWfwQdnoiIiOQIJXApVD21jOKiAgZ17Oanxbdy3IY6mHeNl7wVqqpFREQkNZRVpFBVZSn3XVrJhEeuZkzrKjj7h3DCNUGHJSIiIjlGCVyKHfPiV6HtXfjn+2DGGUGHIyIiIjlICVyqffinsL8Nxh0RdCQiIiKSo5TApdqY6UFHICIiIjlO68CJiIiIhIwSOBEREZGQUQInIiIiEjJK4ERERERCRgmciIiISMgogRMREREJGSVwIiIiIiGjBE5EREQkZJTAiYiIiISMEjgRERGRkFECJyIiIhIySuBEREREQkYJnIiIiEjIKIETERERCRklcCIiIiIhY865oGPIKDPbCjSk8JZjgG0pvF/YqT66Un10Ul2khuqxK9VH36nOusr2+qh0zpV3L8y7BC7VzOxl59zcoOPIFqqPrlQfnVQXqaF67Er10Xeqs67CWh/qQhUREREJGSVwIiIiIiGjBG7gbgs6gCyj+uhK9dFJdZEaqseuVB99pzrrKpT1oTFwIiIiIiGjFjgRERGRkMm5BM7MJpnZEjN708zeMLMv+eWjzWyxma3yn0v98svMbJmZvW5mz5nZ7Jh73W5mjWa2vJfvebaZrTSzejO7IaZ8kV++3L/XoB6uP8zMXvCvv9fMiv3yU83sFTPrMLOL8qg+rvOvdWY2JqZ8vpntNLOl/uObIa6L35rZa/79/2xmw3q4vsr//vVm9nMzM7/8E/7PEDGzfs2eCml93Gxm682spVv51Wa2Nea9saA/ddIf2VSPMcd/3r2Ouh3Pi/dVH+oj0PdVNtWZmd1pZu/E/MzH9nD9YZYfn1vJ1kdaPrcScs7l1AOYAMzxvx4OvA0cAfwAuMEvvwH4vv/1SUCp//U5wAsx9zoVmAMsT/D9CoHVwFSgGHgNOMI/di5g/uMPwOd7uMd9wCX+17+KngdMAY4B7gIuyqP6OM7/2dcCY2LK5wMP58h7Y0TMef8V/f5x7vEiUO3X2aPAOX754cBMoAaYm0f1Ue3H3dKt/Grglv6+NwbyyKZ69I/PBe7uXkf5+L7qQ30E+r7KpjoD7iSJzxvy53Mr2fpIy+dWwu+Z7jdm0A/gQeAMYCUwIebNsTLOuaXAxm5lU3r5hz8ReDzm9Y3AjXHO+z/AzXHKDW8BwaJ49+vLGygX6qPbOWn9RciGuvD//X8JfC3O9ROAFTGvLwV+3e2cGvr5QRu2+uh2XtYkcNlUj3gfREuIk4zk4/sqmfrIxvdVwHV2J7183pBHn1t9/TnIYAKXc12oscxsCl5W/AIwzjm3yT+0GRgX55JP4/1vtC8mAutjXm/wy2LjGARcATwW5/oyYIdzrqOn61MlJPXRmxPN62p71MyO7Mf10RimEHBdmNkd/vebBfx3D9dv6On6VApJffTm49bZBTupH9cPWBbU43XAQzHft6fr8+V9lUx99Caj76ssqDOAm/2f+SdmVhLn+nz73OqtPnqTks+t7nI2gTNvDM1fgH91zu2KPea8tNh1O/80vH/4r6UhnF8ATzvnnknDvZOSI/XxCt6WIrPxPuAf6M83z5a6cM59EjgUeAu4OJX37oscqY+/AVOcc8cAi4HfpTK2ZARdj2Z2KPAJ+pf8plyO1EdG31dB15nvRrz/RB0PjE7xvfskR+ojJZ9b8eRkAue38PwFWOSc+6tfvMXMJvjHJwCNMecfAywELnDONfVy70kxgxE/B2wEYv9XVuGXRc//FlAOXB9T9rh//UKgCRhlZkXxrk+FkNVHj5xzu5xzLf7XjwCDYgeLJiOb6sL/OfYDf8T7X35hzPU3+edWJLp+oEJWHz1yzjU55/b5LxcCVYl/8tTKkno8DpgO1JvZWuAQf1B1vr6vkq2PHmXyfZUldYZzbpPz7APuAOb598jLz60k66NHqfjcSnTznHrg9c3fBfy0W/kP6Tr48Qf+15OBeuCkHu43hcR950XAGuAwOgc/HukfWwA8BwzpJeY/0XUw6Be6Hb+T/g8GDV19xNxrLV3HEoyHA2sXzgPWRV+HqS78OKbHxPQj4Ec93KP7YPNzux2vof+DzUNXHzH36j5WaULM1x8FavtTJ2Gux97qKN/eV32pj6DfV9lUZ3SOMTPgp8B/9nCPfPncSqo+Yu61lhR+biX8Xul4Mwb5AE7Ba1ZdBiz1H+fi9dn/HVgFPAmM9s9fCDTHnPtyzL3+AGwC2vH6xD/dw/c8F2+WzGrg6zHlHX5Z9N7f7OH6qXh/UOv9X4oSv/x4//u24v2P5408qY8v+vfvAN4FFvrl1wFv+L9ctT39smZ7XeC1fD8LvA4sBxYRMwuz2/Vz/XNWA7fQ+Yfgo/733QdsodsA4hyujx/494/4z9/2y/8j5r2xBJiVj39zup2TKIHL+fdVH+sj0PdVNtUZ8BSdv4v3AMN6uD5fPreSrY+0fG4lemgnBhEREZGQyckxcCIiIiK5TAmciIiISMgogRMREREJGSVwIiIiIiGjBE5EREQkZJTAiYh0Y2ZlMQt9bjazjf7XLWb2i6DjExHRMiIiIgmY2bfx1hD7UdCxiIhEqQVORCRJZjbfzB72v/62mf3OzJ4xswYz+5iZ/cDMXjezx/ytgDCzKjP7XzOr87ffmRDsTyEiuUAJnIhI/00DTgfOx1ulfYlz7mhgD3Cen8T9N96WQlXA7cDNQQUrIrmjqPdTRESkB48659rN7HWgEHjML38db//FmcBRwGIzwz9nUwBxikiOUQInItJ/+wCccxEza3edg4ojeH9fDW8vyBODClBEcpO6UEVE0mclUG5mJwKY2SAzOzLgmEQkByiBExFJE+dcG3AR8H0zew1YCpwUbFQikgu0jIiIiIhIyKgFTkRERCRklMCJiIiIhIwSOBEREZGQUQInIiIiEjJK4ERERERCRgmciIiISMgogRMREREJGSVwIiIiIiHz/wF4znNKs02NVgAAAABJRU5ErkJggg==\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Model 2: Dense (Window = 30, horizon = 1)"
      ],
      "metadata": {
        "id": "eGdlrr9nI64b"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "HORIZON = 1 # predict one day of Bitcoin Price\n",
        "WINDOW_SIZE = 30 # use 30 timesteps in past"
      ],
      "metadata": {
        "id": "34ANafqbId_3"
      },
      "execution_count": 54,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Make windowed data with appropriate horizon and window sizes\n",
        "full_windows, full_labels = make_windows(prices, window_size = WINDOW_SIZE, horizon = HORIZON)\n",
        "len(full_windows), len(full_labels)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "i4L_7m-2J9Jy",
        "outputId": "a77951d5-76d2-4132-f610-cc30d2ed1234"
      },
      "execution_count": 55,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Window step:\n",
            "  [[ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23\n",
            "  24 25 26 27 28 29 30]]\n",
            "Window indexes:\n",
            " (array([[ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15,\n",
            "        16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30],\n",
            "       [ 1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16,\n",
            "        17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31],\n",
            "       [ 2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16, 17,\n",
            "        18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32]]), array([[2754, 2755, 2756, 2757, 2758, 2759, 2760, 2761, 2762, 2763, 2764,\n",
            "        2765, 2766, 2767, 2768, 2769, 2770, 2771, 2772, 2773, 2774, 2775,\n",
            "        2776, 2777, 2778, 2779, 2780, 2781, 2782, 2783, 2784],\n",
            "       [2755, 2756, 2757, 2758, 2759, 2760, 2761, 2762, 2763, 2764, 2765,\n",
            "        2766, 2767, 2768, 2769, 2770, 2771, 2772, 2773, 2774, 2775, 2776,\n",
            "        2777, 2778, 2779, 2780, 2781, 2782, 2783, 2784, 2785],\n",
            "       [2756, 2757, 2758, 2759, 2760, 2761, 2762, 2763, 2764, 2765, 2766,\n",
            "        2767, 2768, 2769, 2770, 2771, 2772, 2773, 2774, 2775, 2776, 2777,\n",
            "        2778, 2779, 2780, 2781, 2782, 2783, 2784, 2785, 2786]]), (2757, 31))\n",
            "Windowed array:\n",
            " [[  123.65499   125.455     108.58483 ...   192.75666   197.4\n",
            "    196.02499]\n",
            " [  125.455     108.58483   118.67466 ...   197.4       196.02499\n",
            "    198.04883]\n",
            " [  108.58483   118.67466   121.33866 ...   196.02499   198.04883\n",
            "    198.93233]\n",
            " ...\n",
            " [63346.78904 61965.7826  60574.44473 ... 49764.13208 50032.69314\n",
            "  47885.62525]\n",
            " [61965.7826  60574.44473 56850.83017 ... 50032.69314 47885.62525\n",
            "  45604.61575]\n",
            " [60574.44473 56850.83017 56224.10159 ... 47885.62525 45604.61575\n",
            "  43144.47129]]\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(2757, 2757)"
            ]
          },
          "metadata": {},
          "execution_count": 55
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Make train and test windows\n",
        "train_windows, test_windows, train_labels, test_labels = make_train_test_splits(windows= full_windows, \n",
        "                                                                                labels = full_labels,\n",
        "                                                                                test_split = 0.2)\n",
        "\n",
        "len(train_windows), len(test_windows), len(train_labels), len(test_labels) "
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FsQ4P-rzKHBk",
        "outputId": "cbb53c10-b7fc-4c28-8ccf-80dc188b28d8"
      },
      "execution_count": 56,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(2205, 552, 2205, 552)"
            ]
          },
          "metadata": {},
          "execution_count": 56
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from traitlets.traitlets import validate\n",
        "# Creating, compiling and fitting model\n",
        "\n",
        "tf.random.set_seed(42)\n",
        "\n",
        "# Create model\n",
        "model_2 = tf.keras.Sequential([\n",
        "    layers.Dense(128, activation = \"relu\"),\n",
        "    layers.Dense(HORIZON, activation = \"linear\") \n",
        "    ], name = \"model_2_dense\")\n",
        "\n",
        "# Compile model\n",
        "model_2.compile(loss = \"mae\",\n",
        "                optimizer = tf.keras.optimizers.Adam())\n",
        "# Fit model\n",
        "\n",
        "model_2.fit(train_windows, \n",
        "            train_labels,\n",
        "            epochs = 100,\n",
        "            batch_size = 128,\n",
        "            verbose = 0,\n",
        "            validation_data=(test_windows, test_labels),\n",
        "            callbacks = [create_model_checkpoint(model_name = model_2.name)])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mqQsq5e5LF6Q",
        "outputId": "ba08edc0-b7dc-4849-df5a-db581291e69f"
      },
      "execution_count": 57,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/assets\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7fc69e1eda10>"
            ]
          },
          "metadata": {},
          "execution_count": 57
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Evaluate model 2 on test data\n",
        "\n",
        "model_2.evaluate(test_windows, test_labels)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QJ0ltkj-Lq2Y",
        "outputId": "cc486eb1-984f-440f-da39-8448da15f505"
      },
      "execution_count": 58,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "18/18 [==============================] - 0s 2ms/step - loss: 659.3332\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "659.3331909179688"
            ]
          },
          "metadata": {},
          "execution_count": 58
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model_2 = tf.keras.models.load_model(\"/content/drive/MyDrive/BTC Predictor/model_experiments/model_2_dense/\")\n",
        "model_2.evaluate(test_windows, test_labels)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LJ3PyVyiQbeV",
        "outputId": "2f47d216-d621-43d1-9787-8fef50bf27ef"
      },
      "execution_count": 59,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "18/18 [==============================] - 0s 2ms/step - loss: 605.8741\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "605.8740844726562"
            ]
          },
          "metadata": {},
          "execution_count": 59
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Get forecast predictions\n",
        "model_2_preds = make_preds(model_2, \n",
        "                           input_data = test_windows)"
      ],
      "metadata": {
        "id": "qgsQ_iQrQsQQ"
      },
      "execution_count": 60,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Evaluate results for model_2 predictions\n",
        "model_2_results = evaluate_preds(y_true = tf.squeeze(test_labels), \n",
        "                                 y_pred = model_2_preds)\n",
        "\n",
        "model_2_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3mkFoLrgQ9b_",
        "outputId": "afb13abe-9d14-46de-d124-ac0af1a0b11e"
      },
      "execution_count": 61,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'mae': 605.8741,\n",
              " 'mape': 2.726445,\n",
              " 'mase': 1.0590738,\n",
              " 'mse': 1273360.8,\n",
              " 'rmse': 1128.4329}"
            ]
          },
          "metadata": {},
          "execution_count": 61
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model_1_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DtDIkhAqRcmC",
        "outputId": "e87f4042-7f32-4f42-f88d-1a42e320447f"
      },
      "execution_count": 62,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'mae': 568.95105,\n",
              " 'mape': 2.5448983,\n",
              " 'mase': 0.9994894,\n",
              " 'mse': 1171743.8,\n",
              " 'rmse': 1082.4711}"
            ]
          },
          "metadata": {},
          "execution_count": 62
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Plotting\n",
        "\n",
        "offset = 300\n",
        "plt.figure(figsize=(10,7))\n",
        "# Account for test_windows offset when plotting\n",
        "plot_time_series(timesteps = X_test[-len(test_windows):], \n",
        "                 values = test_labels[:, 0], \n",
        "                 start = offset, \n",
        "                 label = \"Test Data\")\n",
        "\n",
        "plot_time_series(timesteps = X_test[-len(test_windows):], \n",
        "                 values = model_2_preds,\n",
        "                 format = '-', \n",
        "                 start = offset,\n",
        "                 label = \"model_2_preds\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 442
        },
        "id": "PBPwF0c8RtGg",
        "outputId": "75083265-ad4b-4b58-f6aa-3d4775450513"
      },
      "execution_count": 63,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x504 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Model 3: Dense ( window = 30, horizon = 7)"
      ],
      "metadata": {
        "id": "-jcVoVG2ZGaE"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "HORIZON = 7\n",
        "WINDOW_SIZE = 30 "
      ],
      "metadata": {
        "id": "Q57AerNtSmed"
      },
      "execution_count": 64,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "prices[:5]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7Lf_YDVZZl4K",
        "outputId": "90b3ab15-884b-479c-ade6-64faf9c86617"
      },
      "execution_count": 65,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([123.65499, 125.455  , 108.58483, 118.67466, 121.33866])"
            ]
          },
          "metadata": {},
          "execution_count": 65
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "full_windows, full_labels = make_windows(prices, window_size = WINDOW_SIZE, horizon = HORIZON)\n",
        "len(full_windows), len(full_labels)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ObiDZpXSZRNm",
        "outputId": "c5bf8cfe-1280-4d6a-d18d-4e386ff86e19"
      },
      "execution_count": 66,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Window step:\n",
            "  [[ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23\n",
            "  24 25 26 27 28 29 30 31 32 33 34 35 36]]\n",
            "Window indexes:\n",
            " (array([[ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15,\n",
            "        16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31,\n",
            "        32, 33, 34, 35, 36],\n",
            "       [ 1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16,\n",
            "        17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32,\n",
            "        33, 34, 35, 36, 37],\n",
            "       [ 2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16, 17,\n",
            "        18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33,\n",
            "        34, 35, 36, 37, 38]]), array([[2748, 2749, 2750, 2751, 2752, 2753, 2754, 2755, 2756, 2757, 2758,\n",
            "        2759, 2760, 2761, 2762, 2763, 2764, 2765, 2766, 2767, 2768, 2769,\n",
            "        2770, 2771, 2772, 2773, 2774, 2775, 2776, 2777, 2778, 2779, 2780,\n",
            "        2781, 2782, 2783, 2784],\n",
            "       [2749, 2750, 2751, 2752, 2753, 2754, 2755, 2756, 2757, 2758, 2759,\n",
            "        2760, 2761, 2762, 2763, 2764, 2765, 2766, 2767, 2768, 2769, 2770,\n",
            "        2771, 2772, 2773, 2774, 2775, 2776, 2777, 2778, 2779, 2780, 2781,\n",
            "        2782, 2783, 2784, 2785],\n",
            "       [2750, 2751, 2752, 2753, 2754, 2755, 2756, 2757, 2758, 2759, 2760,\n",
            "        2761, 2762, 2763, 2764, 2765, 2766, 2767, 2768, 2769, 2770, 2771,\n",
            "        2772, 2773, 2774, 2775, 2776, 2777, 2778, 2779, 2780, 2781, 2782,\n",
            "        2783, 2784, 2785, 2786]]), (2751, 37))\n",
            "Windowed array:\n",
            " [[  123.65499   125.455     108.58483 ...   210.3075    225.02\n",
            "    248.25333]\n",
            " [  125.455     108.58483   118.67466 ...   225.02      248.25333\n",
            "    262.32666]\n",
            " [  108.58483   118.67466   121.33866 ...   248.25333   262.32666\n",
            "    294.48699]\n",
            " ...\n",
            " [58171.90902 59295.95004 59822.90168 ... 49764.13208 50032.69314\n",
            "  47885.62525]\n",
            " [59295.95004 59822.90168 59853.19724 ... 50032.69314 47885.62525\n",
            "  45604.61575]\n",
            " [59822.90168 59853.19724 63223.88439 ... 47885.62525 45604.61575\n",
            "  43144.47129]]\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(2751, 2751)"
            ]
          },
          "metadata": {},
          "execution_count": 66
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train_windows, test_windows, train_labels, test_labels = make_train_test_splits(windows = full_windows, labels = full_labels,\n",
        "                                                                                 test_split = 0.2) \n",
        "len(train_windows), len(test_windows), len(train_labels), len(test_labels)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "P5X8ABOaZn7h",
        "outputId": "1eca88c4-83b0-47ed-eb61-1e0ed6d2f97e"
      },
      "execution_count": 67,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(2200, 551, 2200, 551)"
            ]
          },
          "metadata": {},
          "execution_count": 67
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "tf.random.set_seed(42)\n",
        "\n",
        "# Create model (same as model 1 except with different data input and output sizes)\n",
        "\n",
        "model_3 = tf.keras.Sequential([ \n",
        "    layers.Dense(128, activation = \"relu\"),\n",
        "    layers.Dense(HORIZON)], \n",
        "    name = \"model_3_dense\")\n",
        "\n",
        "# Compiling model\n",
        "\n",
        "model_3.compile(loss = \"mae\",\n",
        "                optimizer = tf.keras.optimizers.Adam())\n",
        "\n",
        "# Fitting model\n",
        "\n",
        "model_3.fit(x=train_windows, y=train_labels,\n",
        "            epochs=100,\n",
        "            batch_size = 128,\n",
        "            verbose = 0,\n",
        "            validation_data = (test_windows, test_labels),\n",
        "            callbacks = [create_model_checkpoint(model_name = model_3.name)])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4atLh80AaiK1",
        "outputId": "f133046a-8ec7-49b7-d1e3-4c3bd869cc17"
      },
      "execution_count": 68,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/assets\n",
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/assets\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7fc679f03dd0>"
            ]
          },
          "metadata": {},
          "execution_count": 68
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Evaluate the model on the test data\n",
        "\n",
        "model_3.evaluate(test_windows, test_labels)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dBFFkr0y4NRL",
        "outputId": "726e68c0-5e03-41f8-ce52-2fdb796b3292"
      },
      "execution_count": 69,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "18/18 [==============================] - 0s 2ms/step - loss: 1300.1525\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "1300.1524658203125"
            ]
          },
          "metadata": {},
          "execution_count": 69
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Load the best version of model 3 and evaluate\n",
        "\n",
        "model_3 = tf.keras.models.load_model(\"/content/drive/MyDrive/BTC Predictor/model_experiments/model_3_dense/\")\n",
        "\n",
        "model_3.evaluate(test_windows, test_labels)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "x7lNJ7Ha6QEb",
        "outputId": "e339c040-ef39-4f61-ede9-17c5897c8f0f"
      },
      "execution_count": 70,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "18/18 [==============================] - 0s 2ms/step - loss: 1237.5065\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "1237.5064697265625"
            ]
          },
          "metadata": {},
          "execution_count": 70
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Making predictions with model 3\n",
        "\n",
        "model_3_preds = make_preds(model_3, \n",
        "                           input_data = test_windows)\n",
        "\n",
        "model_3_preds[:5]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "V4zz9Bll6y_8",
        "outputId": "14a04f0f-db14-4dda-d98d-3766fcd828f3"
      },
      "execution_count": 71,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Tensor: shape=(5, 7), dtype=float32, numpy=\n",
              "array([[9004.694 , 9048.1   , 9425.09  , 9258.256 , 9495.797 , 9558.453 ,\n",
              "        9357.3545],\n",
              "       [8735.506 , 8840.303 , 9247.793 , 8885.601 , 9097.186 , 9174.327 ,\n",
              "        9156.818 ],\n",
              "       [8672.508 , 8782.387 , 9123.857 , 8770.371 , 9007.13  , 9003.869 ,\n",
              "        9042.725 ],\n",
              "       [8874.399 , 8784.74  , 9043.902 , 8943.052 , 9033.477 , 9176.489 ,\n",
              "        9039.677 ],\n",
              "       [8825.891 , 8777.44  , 8926.78  , 8870.179 , 9213.232 , 9268.156 ,\n",
              "        8942.487 ]], dtype=float32)>"
            ]
          },
          "metadata": {},
          "execution_count": 71
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model_2_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_nVyHlXe8Xc0",
        "outputId": "5d7f163f-f74a-479d-e173-ac61747b60bf"
      },
      "execution_count": 72,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'mae': 605.8741,\n",
              " 'mape': 2.726445,\n",
              " 'mase': 1.0590738,\n",
              " 'mse': 1273360.8,\n",
              " 'rmse': 1128.4329}"
            ]
          },
          "metadata": {},
          "execution_count": 72
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model_3_preds.shape, model_2_preds.shape, model_1_preds.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gxOMdHKe8iyi",
        "outputId": "45d75fe2-8dcc-48c1-e6bd-c9abb77ccc1f"
      },
      "execution_count": 73,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(TensorShape([551, 7]), TensorShape([552]), TensorShape([556]))"
            ]
          },
          "metadata": {},
          "execution_count": 73
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Evaluate model_3_results\n",
        "\n",
        "model_3_results = evaluate_preds( y_true = tf.squeeze(test_labels), \n",
        "                                 y_pred = model_3_preds)\n",
        "\n",
        "model_3_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "jk6hgiwJ7BO6",
        "outputId": "f9b34db2-5bed-45b1-8ecc-6f60c3a627c9"
      },
      "execution_count": 74,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'mae': array([ 513.6056  ,  355.08328 ,  327.17075 ,  358.5106  ,  420.53305 ,\n",
              "         537.854   ,  545.66156 ,  485.9228  ,  584.49713 ,  687.3824  ,\n",
              "         836.2264  ,  755.1577  ,  731.4958  ,  775.3399  ,  567.95544 ,\n",
              "         266.80865 ,  188.80225 ,  188.10799 ,  253.095   ,  301.43256 ,\n",
              "         151.10721 ,  196.81473 ,  191.46178 ,  231.65031 ,  143.61125 ,\n",
              "         122.59089 ,  132.78879 ,  190.8118  ,  179.15994 ,  228.26018 ,\n",
              "         314.44022 ,  379.09244 ,  278.32526 ,  295.3471  ,  299.3852  ,\n",
              "         248.64983 ,  299.75662 ,  259.69434 ,  180.30531 ,  206.72922 ,\n",
              "         374.62863 ,  144.85086 ,  142.33635 ,  131.11613 ,   93.941124,\n",
              "          54.825542,   73.79541 ,  103.60128 ,  121.333916,  168.67278 ,\n",
              "         183.90945 ,  152.25307 ,  186.57191 ,  146.91267 ,  240.42912 ,\n",
              "         351.00662 ,  540.952   ,  549.1573  ,  521.2411  ,  526.8554  ,\n",
              "         453.36237 ,  257.98117 ,  277.29486 ,  301.8247  ,  455.718   ,\n",
              "         458.96017 ,  503.4438  ,  522.3119  ,  223.07631 ,  250.09529 ,\n",
              "         297.1437  ,  400.571   ,  495.7984  ,  364.33707 ,  283.36606 ,\n",
              "         325.59543 ,  238.21191 ,  318.9777  ,  460.77106 ,  651.0762  ,\n",
              "         835.88    ,  669.96454 ,  319.54507 ,  261.99442 ,  142.3923  ,\n",
              "         136.62863 ,  154.75223 ,  221.32631 ,  290.50418 ,  503.88406 ,\n",
              "         414.2602  ,  434.35742 ,  377.988   ,  251.78947 ,  204.2832  ,\n",
              "         388.227   ,  360.65793 ,  493.80902 ,  614.861   ,  754.70245 ,\n",
              "         533.7712  ,  378.98898 ,  280.49496 ,  339.48062 ,  413.12875 ,\n",
              "         452.87863 ,  550.53796 ,  634.5724  ,  935.5724  ,  931.60004 ,\n",
              "         881.2808  ,  426.40012 ,  179.45871 ,  121.66183 ,  160.43861 ,\n",
              "         372.07108 ,  341.8582  ,  476.52377 ,  618.32434 , 1038.8987  ,\n",
              "        1569.5028  , 2157.1199  , 1987.6074  , 2158.6108  , 2303.561   ,\n",
              "        2662.943   , 1405.5012  ,  728.30096 ,  351.70068 ,  322.032   ,\n",
              "         493.94748 ,  435.48416 ,  565.556   ,  350.11685 ,  289.01987 ,\n",
              "         251.21603 ,  409.05615 ,  342.2099  ,  320.83615 ,  330.88617 ,\n",
              "         357.84604 ,  335.94818 ,  206.10303 ,  544.38416 ,  700.00934 ,\n",
              "         468.91992 ,  404.49615 ,  172.80127 ,  308.33685 ,  210.47621 ,\n",
              "         318.9543  ,  486.1326  ,  428.8803  ,  533.9111  ,  433.78146 ,\n",
              "         396.89432 ,  138.40381 ,  189.96658 ,  170.39209 ,  181.54346 ,\n",
              "         282.88992 ,  264.38867 ,  250.6218  ,  240.33762 ,  276.95425 ,\n",
              "         326.02963 ,  489.6155  ,  686.2448  ,  526.9793  ,  603.01135 ,\n",
              "         825.3831  ,  871.0465  ,  990.0687  , 1090.0597  ,  560.2842  ,\n",
              "         310.52246 ,  371.39035 ,  348.45953 ,  355.73438 ,  429.56543 ,\n",
              "         581.2835  ,  550.52203 ,  635.4312  ,  913.404   ,  840.7183  ,\n",
              "         305.0357  ,  493.93332 ,  751.31793 ,  410.6346  ,  220.62444 ,\n",
              "         282.25446 ,  291.8528  ,  422.5039  ,  458.6543  ,  637.2352  ,\n",
              "         647.82336 ,  417.24496 ,  220.23703 ,  246.72209 ,  200.22922 ,\n",
              "         455.1494  ,  719.6059  ,  696.3024  ,  485.0996  ,  294.95297 ,\n",
              "         170.52371 ,  211.82478 ,  270.565   ,  189.89299 ,  171.21736 ,\n",
              "         366.44727 ,  231.96387 ,  318.78738 ,  273.79395 ,  358.55664 ,\n",
              "         412.22797 ,  512.3167  ,  185.43959 ,  196.38225 ,  200.006   ,\n",
              "         224.58678 ,  213.41978 ,  186.23912 ,  113.371376,  172.23228 ,\n",
              "         168.89857 ,  236.0957  ,  307.5971  ,  328.93903 ,  566.59625 ,\n",
              "         285.04395 ,  300.44965 ,  125.45299 ,  168.94322 ,  137.25851 ,\n",
              "         143.5039  ,  145.27776 ,  107.339424,   77.16225 ,  131.87082 ,\n",
              "         134.02023 ,  167.45773 ,  137.9216  ,  148.91336 ,  204.95508 ,\n",
              "         157.89746 ,  196.95256 ,  167.93918 ,  156.45605 ,  188.68094 ,\n",
              "         161.44127 ,   90.99735 ,  136.8411  ,  198.51701 ,  230.1381  ,\n",
              "         294.78348 ,  594.2425  ,  699.6479  ,  815.13574 ,  905.339   ,\n",
              "        1127.2997  , 1342.6945  , 1317.0691  ,  590.1681  ,  296.34515 ,\n",
              "         243.43625 ,  256.3983  ,  222.28459 ,  323.83832 ,   82.513115,\n",
              "         120.14384 ,  249.14914 ,  205.73592 ,  243.4548  ,  250.64803 ,\n",
              "         287.2761  ,  224.38142 ,  266.56027 ,  221.51047 ,  218.4965  ,\n",
              "         272.42285 ,  279.59625 ,  252.5883  ,  381.565   ,  417.97086 ,\n",
              "         624.5573  ,  368.03738 ,  327.25488 ,  263.43973 ,  349.95297 ,\n",
              "         398.62527 ,  297.07407 ,  147.04924 ,  164.5431  ,  313.36328 ,\n",
              "         477.7482  ,  675.0409  ,  897.2679  , 1094.7103  , 1460.1775  ,\n",
              "        1398.586   ,  952.72504 ,  645.959   ,  166.17899 ,  144.66895 ,\n",
              "         189.18289 ,  304.8132  ,  435.10602 ,  449.64383 ,  425.2437  ,\n",
              "         441.93597 ,  407.29465 ,  252.40388 ,  248.36984 ,  336.17105 ,\n",
              "         482.8107  ,  437.53098 ,  533.0868  ,  346.97934 ,  127.68666 ,\n",
              "         110.20843 ,  301.75113 ,  195.50545 ,  174.65164 ,  238.47084 ,\n",
              "         302.3718  ,  313.51758 ,  310.1296  ,  200.29842 ,  172.47307 ,\n",
              "         140.2302  ,  252.48549 ,  289.32437 ,  343.62094 ,  504.1179  ,\n",
              "         635.1342  ,  602.13043 ,  519.06055 ,  214.90276 ,  195.9117  ,\n",
              "         265.0371  ,  198.90123 ,  345.51285 ,  517.2228  ,  631.0296  ,\n",
              "         988.4247  , 1174.214   , 1196.2844  , 1253.9299  ,  526.0653  ,\n",
              "         210.31487 ,  215.27107 ,  169.86008 ,  283.78656 ,  269.76157 ,\n",
              "         228.6349  ,  186.64648 ,  410.54242 ,  601.2651  ,  618.1225  ,\n",
              "         768.7529  , 1158.5457  , 1232.8796  , 1254.4298  ,  423.0244  ,\n",
              "         390.0357  ,  367.23856 ,  209.55775 ,  530.22235 ,  821.38226 ,\n",
              "         812.3729  ,  741.1964  ,  953.4834  , 1258.9906  , 1844.932   ,\n",
              "        1605.2845  , 1112.7659  ,  594.18835 ,  549.7667  ,  632.5435  ,\n",
              "         829.2642  , 1103.2218  , 1130.1029  , 1033.3522  ,  878.5865  ,\n",
              "         595.4085  , 1115.6506  , 1371.1725  , 1385.2511  ,  387.5229  ,\n",
              "         303.7486  ,  495.125   ,  719.5804  ,  648.90265 ,  766.1632  ,\n",
              "         683.23157 ,  596.4121  ,  523.09265 ,  577.3407  , 1337.6213  ,\n",
              "        2454.5972  , 2759.2344  , 3135.8762  , 3407.3806  , 3602.6335  ,\n",
              "        2527.1577  , 1158.0999  ,  679.8432  ,  748.5586  , 1073.4863  ,\n",
              "        1217.7327  , 1459.5656  , 2502.731   , 3075.8247  , 3090.2878  ,\n",
              "        2564.7583  , 2660.5317  , 2928.934   , 3690.556   , 3525.1406  ,\n",
              "        4183.6694  , 5144.068   , 4384.655   , 4164.9604  , 4431.0938  ,\n",
              "        3335.1216  , 3017.6858  , 2589.5396  , 4103.63    , 5582.0933  ,\n",
              "        5108.46    , 1382.9637  ,  953.5419  , 1081.6825  , 2483.1975  ,\n",
              "        2992.0547  , 3127.0942  , 2972.274   , 3054.65    , 3283.713   ,\n",
              "        3617.3674  , 1170.2589  , 1074.1833  , 1059.317   , 1044.5404  ,\n",
              "        1060.6217  , 2115.719   , 3569.8875  , 3474.363   , 2448.0156  ,\n",
              "        2641.418   , 3188.6028  , 4899.915   , 5516.288   , 4635.486   ,\n",
              "        4677.034   , 5732.883   , 5866.738   , 8083.3403  , 5049.8286  ,\n",
              "        1476.2288  , 1873.3275  , 1219.0994  , 2324.073   , 2467.0095  ,\n",
              "        3005.0898  , 2805.0503  , 3688.254   , 2962.7136  , 3664.2388  ,\n",
              "        5618.684   , 6925.097   , 9751.095   , 8790.821   , 5057.626   ,\n",
              "        2971.6848  , 1715.495   , 1125.7706  , 2201.4253  , 3969.4978  ,\n",
              "        2988.5083  , 2911.9285  , 2796.8672  , 3937.8806  , 5449.5894  ,\n",
              "        6395.021   , 6148.243   , 5437.6997  , 4291.8267  , 2362.3967  ,\n",
              "        1657.3405  , 1426.6367  , 1647.7155  , 2730.398   , 1813.7115  ,\n",
              "        2061.1506  , 3095.7227  , 4312.8096  , 5500.727   , 5258.027   ,\n",
              "        5152.0283  , 1331.4062  , 1634.4023  , 2493.4324  , 3957.9526  ,\n",
              "        4499.2144  , 2215.846   , 2587.4119  , 1622.4397  ,  591.875   ,\n",
              "        1297.4727  , 1284.5977  ,  915.22766 ,  940.44977 , 1084.6584  ,\n",
              "         875.8979  , 1041.3549  , 2299.2473  , 3152.6003  , 3518.394   ,\n",
              "        2531.524   , 2682.5737  , 3094.2917  , 3829.3372  , 5550.616   ,\n",
              "        7159.3477  , 8820.524   , 8587.342   , 6777.815   , 5193.292   ,\n",
              "        4793.3384  , 2605.2556  , 1668.3605  , 2510.7183  , 4105.6826  ,\n",
              "        6278.394   , 4109.754   , 1711.1172  , 2261.9832  , 2384.7534  ,\n",
              "        1246.6641  , 1672.8483  , 1103.1345  , 1743.3678  , 1608.4564  ,\n",
              "        1933.904   , 2742.3433  , 3945.3645  , 5765.941   , 6955.8433  ,\n",
              "        8107.7095  ], dtype=float32),\n",
              " 'mape': array([ 5.860151  ,  4.087341  ,  3.775816  ,  4.187658  ,  4.9781423 ,\n",
              "         6.448365  ,  6.6144614 ,  6.061672  ,  7.5475917 ,  9.095927  ,\n",
              "        11.300206  , 10.396474  , 10.109089  , 10.697326  ,  7.8628473 ,\n",
              "         3.6872826 ,  2.585637  ,  2.505004  ,  3.350452  ,  4.013581  ,\n",
              "         2.0036225 ,  2.6397398 ,  2.6087885 ,  3.1609263 ,  1.9612128 ,\n",
              "         1.6679873 ,  1.7952291 ,  2.5901227 ,  2.4362576 ,  3.1327405 ,\n",
              "         4.3595424 ,  5.285637  ,  3.940355  ,  4.3142257 ,  4.3733144 ,\n",
              "         3.6376853 ,  4.3629785 ,  3.7757437 ,  2.6192946 ,  2.900764  ,\n",
              "         5.1501083 ,  2.0147362 ,  1.9706653 ,  1.8104146 ,  1.3007245 ,\n",
              "         0.7535146 ,  1.0192324 ,  1.4282739 ,  1.6741855 ,  2.3589702 ,\n",
              "         2.563099  ,  2.1327195 ,  2.5974853 ,  2.0173264 ,  3.1387615 ,\n",
              "         4.502132  ,  6.95963   ,  6.9602227 ,  6.5331383 ,  6.5757856 ,\n",
              "         5.610285  ,  3.0642037 ,  3.2474003 ,  3.484728  ,  5.2227206 ,\n",
              "         5.2294765 ,  5.727937  ,  5.927383  ,  2.5284474 ,  2.8664036 ,\n",
              "         3.4308906 ,  4.68742   ,  5.841512  ,  4.309815  ,  3.3552825 ,\n",
              "         3.8162112 ,  2.7363598 ,  3.4623704 ,  4.9716797 ,  7.0261016 ,\n",
              "         8.974979  ,  7.1678147 ,  3.4013336 ,  2.7761197 ,  1.5223246 ,\n",
              "         1.4535493 ,  1.6354663 ,  2.252781  ,  2.947562  ,  5.0738955 ,\n",
              "         4.103283  ,  4.2952995 ,  3.7052832 ,  2.4620771 ,  2.0132227 ,\n",
              "         3.84283   ,  3.6306522 ,  5.016276  ,  6.279475  ,  7.757472  ,\n",
              "         5.495675  ,  3.8907118 ,  2.854692  ,  3.531113  ,  4.464891  ,\n",
              "         5.01624   ,  6.1257133 ,  7.1294346 , 10.711447  , 10.664077  ,\n",
              "        10.087252  ,  4.8936324 ,  2.0717275 ,  1.3628366 ,  1.8011123 ,\n",
              "         4.219881  ,  4.055701  ,  5.7448916 ,  7.4928775 , 15.032983  ,\n",
              "        24.38847   , 35.428123  , 34.702423  , 39.8611    , 42.836586  ,\n",
              "        49.604614  , 26.081118  , 13.669951  ,  6.3608713 ,  5.4986434 ,\n",
              "         7.9925137 ,  6.857536  ,  8.820607  ,  5.311202  ,  4.527914  ,\n",
              "         3.8704064 ,  6.2283874 ,  5.2470293 ,  5.1045513 ,  5.3294277 ,\n",
              "         5.7789717 ,  5.358489  ,  3.2211251 ,  8.118447  , 10.231979  ,\n",
              "         6.726865  ,  5.757307  ,  2.404117  ,  4.2864566 ,  2.9776995 ,\n",
              "         4.4548492 ,  6.9262066 ,  6.1590796 ,  7.763827  ,  6.3082023 ,\n",
              "         5.7371473 ,  1.9719498 ,  2.7191217 ,  2.399884  ,  2.5685852 ,\n",
              "         3.9529998 ,  3.6584747 ,  3.4535587 ,  3.301876  ,  3.794225  ,\n",
              "         4.310312  ,  6.4319797 ,  8.567103  ,  6.293656  ,  7.023875  ,\n",
              "         9.469172  ,  9.915061  , 11.217032  , 12.280664  ,  6.209284  ,\n",
              "         3.2802799 ,  3.8508568 ,  3.6124706 ,  3.7172365 ,  4.6127863 ,\n",
              "         6.233847  ,  6.0376983 ,  7.135962  , 10.132261  ,  9.337389  ,\n",
              "         3.379456  ,  5.190334  ,  7.8229227 ,  4.264503  ,  2.3285017 ,\n",
              "         3.0320892 ,  3.1225486 ,  4.572195  ,  5.01346   ,  7.059326  ,\n",
              "         7.1808734 ,  4.628824  ,  2.4294448 ,  2.6893463 ,  2.1346915 ,\n",
              "         4.670391  ,  7.4296794 ,  7.17774   ,  4.927452  ,  2.9697587 ,\n",
              "         1.7036037 ,  2.102083  ,  2.790086  ,  1.9540966 ,  1.7562671 ,\n",
              "         3.8170424 ,  2.455728  ,  3.3356586 ,  2.8912692 ,  3.8006883 ,\n",
              "         4.368519  ,  5.4371586 ,  1.969062  ,  2.0868657 ,  2.132989  ,\n",
              "         2.4014482 ,  2.2829754 ,  1.9878371 ,  1.212721  ,  1.8306261 ,\n",
              "         1.815668  ,  2.5207548 ,  3.292389  ,  3.5837185 ,  6.187771  ,\n",
              "         3.1184072 ,  3.2884552 ,  1.3740312 ,  1.8472781 ,  1.5017242 ,\n",
              "         1.56714   ,  1.5889224 ,  1.1630667 ,  0.83712757,  1.4223003 ,\n",
              "         1.4407552 ,  1.8029895 ,  1.4796977 ,  1.599085  ,  2.2175858 ,\n",
              "         1.7130921 ,  2.1422663 ,  1.8232723 ,  1.7015604 ,  2.0563033 ,\n",
              "         1.7596658 ,  0.98229367,  1.4560665 ,  2.1070144 ,  2.4114134 ,\n",
              "         3.0499952 ,  5.77384   ,  6.6185617 ,  7.4951463 ,  8.227665  ,\n",
              "        10.186892  , 11.928123  , 11.676128  ,  5.199712  ,  2.592971  ,\n",
              "         2.1249874 ,  2.23352   ,  1.9233074 ,  2.8466196 ,  0.7076641 ,\n",
              "         1.0350558 ,  2.1322331 ,  1.7789838 ,  2.1161134 ,  2.1635559 ,\n",
              "         2.4792871 ,  1.9270623 ,  2.255131  ,  1.845723  ,  1.811268  ,\n",
              "         2.2626755 ,  2.3515134 ,  2.117338  ,  3.2111607 ,  3.5736072 ,\n",
              "         5.371579  ,  3.1869485 ,  2.8585389 ,  2.2958062 ,  3.0580995 ,\n",
              "         3.4806094 ,  2.5900195 ,  1.2625039 ,  1.405033  ,  2.7622333 ,\n",
              "         4.3049507 ,  6.2830634 ,  8.513557  , 10.517092  , 14.161083  ,\n",
              "        13.64061   ,  9.320249  ,  6.3080707 ,  1.6300529 ,  1.3983192 ,\n",
              "         1.819187  ,  2.8875935 ,  4.0688186 ,  4.173644  ,  3.9228578 ,\n",
              "         4.0472693 ,  3.7387352 ,  2.3069754 ,  2.306329  ,  3.194118  ,\n",
              "         4.589591  ,  4.1579466 ,  5.0558167 ,  3.300564  ,  1.2102793 ,\n",
              "         1.0371078 ,  2.8088493 ,  1.8151473 ,  1.6353133 ,  2.2353556 ,\n",
              "         2.8412106 ,  2.9520082 ,  2.918753  ,  1.8913203 ,  1.6178724 ,\n",
              "         1.2982846 ,  2.3039372 ,  2.6034815 ,  3.0176406 ,  4.4442587 ,\n",
              "         5.598941  ,  5.266945  ,  4.535595  ,  1.8671705 ,  1.7055051 ,\n",
              "         2.3196983 ,  1.7318262 ,  2.7897274 ,  4.062607  ,  4.8944464 ,\n",
              "         7.6816387 ,  9.1078005 ,  9.207178  ,  9.502263  ,  3.9754753 ,\n",
              "         1.5744725 ,  1.5996995 ,  1.242373  ,  2.077033  ,  1.9747171 ,\n",
              "         1.6764205 ,  1.3502657 ,  2.785834  ,  3.9915614 ,  4.0835915 ,\n",
              "         5.016293  ,  7.6388426 ,  8.078356  ,  8.109144  ,  2.6612601 ,\n",
              "         2.4994817 ,  2.296388  ,  1.3040293 ,  3.2585542 ,  4.9145126 ,\n",
              "         4.7377963 ,  4.2538853 ,  5.3531475 ,  6.980067  , 10.170471  ,\n",
              "         8.772198  ,  6.0037904 ,  3.184009  ,  3.0089974 ,  3.6370165 ,\n",
              "         4.774086  ,  6.289237  ,  6.4191914 ,  5.904084  ,  4.996318  ,\n",
              "         3.1905975 ,  5.85623   ,  7.181029  ,  7.2116575 ,  2.0219858 ,\n",
              "         1.5912374 ,  2.6243296 ,  3.8534112 ,  3.5140622 ,  4.1405344 ,\n",
              "         3.6843975 ,  3.215591  ,  2.8228238 ,  2.88985   ,  6.284895  ,\n",
              "        11.458011  , 12.342679  , 13.76467   , 14.824157  , 15.566195  ,\n",
              "        10.829554  ,  4.9278164 ,  2.8345594 ,  3.0226746 ,  4.3117876 ,\n",
              "         4.720057  ,  5.5167937 ,  9.303037  , 11.212705  , 11.037068  ,\n",
              "         8.631595  ,  8.57773   ,  9.389047  , 11.49777   , 10.463528  ,\n",
              "        11.683611  , 13.823773  , 11.273861  , 10.805959  , 11.400708  ,\n",
              "         8.496614  ,  7.8893795 ,  7.1694746 , 11.423845  , 15.540586  ,\n",
              "        14.24639   ,  3.8430889 ,  2.5450244 ,  2.926231  ,  7.2883105 ,\n",
              "         8.999361  ,  9.586315  ,  9.271843  ,  9.42326   , 10.242951  ,\n",
              "        11.39661   ,  3.655802  ,  3.3454576 ,  3.2290843 ,  3.1591043 ,\n",
              "         3.1852615 ,  6.141231  , 10.163145  ,  9.673941  ,  6.657334  ,\n",
              "         6.975822  ,  8.377457  , 12.244591  , 13.002367  , 10.547288  ,\n",
              "        10.269909  , 12.405405  , 12.598534  , 17.187767  , 10.616799  ,\n",
              "         3.0717638 ,  3.766271  ,  2.39153   ,  4.426362  ,  4.6217623 ,\n",
              "         5.4785867 ,  5.115232  ,  6.8264303 ,  5.6151514 ,  7.090962  ,\n",
              "        11.599259  , 14.5101595 , 20.67855   , 18.67078   , 10.813845  ,\n",
              "         6.3959217 ,  3.68131   ,  2.3239684 ,  4.459366  ,  8.030946  ,\n",
              "         5.953289  ,  5.6428604 ,  5.137942  ,  7.162542  ,  9.831411  ,\n",
              "        11.2098    , 10.556696  ,  9.350716  ,  7.3507614 ,  3.9944015 ,\n",
              "         2.82334   ,  2.3929412 ,  2.8540518 ,  4.729089  ,  3.2363744 ,\n",
              "         3.6810775 ,  5.66858   ,  8.005048  , 10.247124  ,  9.747851  ,\n",
              "         9.54604   ,  2.477174  ,  2.9854128 ,  4.34938   ,  6.8806424 ,\n",
              "         7.7586837 ,  3.8011284 ,  4.4251747 ,  2.761091  ,  1.0093703 ,\n",
              "         2.242207  ,  2.2289226 ,  1.5868819 ,  1.6239101 ,  1.8814805 ,\n",
              "         1.5162615 ,  1.7166462 ,  3.7594323 ,  5.0977936 ,  5.66472   ,\n",
              "         4.075433  ,  4.348675  ,  5.1419644 ,  6.618981  ,  9.824003  ,\n",
              "        12.98138   , 16.388012  , 16.260683  , 13.2497835 , 10.180601  ,\n",
              "         9.383696  ,  5.121853  ,  3.2893925 ,  4.634882  ,  7.40683   ,\n",
              "        11.284329  ,  7.2873406 ,  3.019626  ,  4.0061307 ,  4.1726418 ,\n",
              "         2.2337658 ,  3.0054157 ,  1.9702908 ,  3.1211088 ,  2.7774823 ,\n",
              "         3.492497  ,  5.1419334 ,  7.6572046 , 11.470653  , 14.311554  ,\n",
              "        17.280174  ], dtype=float32),\n",
              " 'mase': 2.2020733,\n",
              " 'mse': array([3.15562562e+05, 1.69165047e+05, 1.44132125e+05, 1.76003203e+05,\n",
              "        2.63520375e+05, 3.91517562e+05, 3.99526219e+05, 3.44422156e+05,\n",
              "        4.93892469e+05, 6.88786938e+05, 9.18702875e+05, 8.00988875e+05,\n",
              "        6.35508688e+05, 6.45535812e+05, 3.57740594e+05, 1.03007414e+05,\n",
              "        5.33133203e+04, 4.65832188e+04, 1.10349016e+05, 1.11646961e+05,\n",
              "        4.35359453e+04, 5.45390977e+04, 4.55849648e+04, 7.64883672e+04,\n",
              "        4.03074453e+04, 2.71076816e+04, 1.90270059e+04, 4.55626055e+04,\n",
              "        4.25344805e+04, 6.63250547e+04, 1.16741703e+05, 1.69440500e+05,\n",
              "        1.10487078e+05, 1.59708281e+05, 1.61630891e+05, 1.23861797e+05,\n",
              "        1.50757391e+05, 1.31196844e+05, 5.50280977e+04, 6.28156445e+04,\n",
              "        1.77474219e+05, 2.68842363e+04, 2.53034434e+04, 2.58045059e+04,\n",
              "        1.34594580e+04, 3.88201562e+03, 9.22788672e+03, 1.22403281e+04,\n",
              "        2.20870137e+04, 3.60827500e+04, 4.87961602e+04, 3.74652578e+04,\n",
              "        4.10563438e+04, 4.17418945e+04, 1.10490461e+05, 2.00761469e+05,\n",
              "        3.48149906e+05, 3.58061312e+05, 3.28374031e+05, 2.96505938e+05,\n",
              "        2.33353781e+05, 1.54046719e+05, 1.43171672e+05, 1.62586703e+05,\n",
              "        2.96452750e+05, 2.66880812e+05, 2.94188594e+05, 3.10067562e+05,\n",
              "        6.27178242e+04, 8.63872109e+04, 9.69540391e+04, 2.01815250e+05,\n",
              "        2.97054594e+05, 2.00675078e+05, 1.27305359e+05, 1.32510266e+05,\n",
              "        8.46021953e+04, 1.74816891e+05, 2.95114281e+05, 4.75889719e+05,\n",
              "        7.21893438e+05, 4.62710188e+05, 1.33154703e+05, 8.13637656e+04,\n",
              "        3.43107852e+04, 3.45205547e+04, 5.36855352e+04, 7.42412266e+04,\n",
              "        1.09339273e+05, 2.79682500e+05, 2.19119391e+05, 1.99934375e+05,\n",
              "        1.65902594e+05, 8.16736484e+04, 4.82389180e+04, 1.65301344e+05,\n",
              "        1.81698328e+05, 3.30644938e+05, 4.63626844e+05, 6.75045625e+05,\n",
              "        3.34911812e+05, 1.69344000e+05, 1.03219461e+05, 1.48348000e+05,\n",
              "        2.98685469e+05, 3.71952281e+05, 4.43305750e+05, 5.27904938e+05,\n",
              "        1.12945225e+06, 1.00652912e+06, 8.03247688e+05, 2.21277750e+05,\n",
              "        6.64889453e+04, 2.66178418e+04, 3.29691094e+04, 1.68663719e+05,\n",
              "        1.76257750e+05, 2.99364688e+05, 4.31794688e+05, 1.91346975e+06,\n",
              "        3.90354150e+06, 6.29502550e+06, 5.49042450e+06, 6.25677100e+06,\n",
              "        6.28618550e+06, 7.23876850e+06, 2.08143425e+06, 6.66785812e+05,\n",
              "        1.44640391e+05, 1.83413781e+05, 3.41026000e+05, 3.02068438e+05,\n",
              "        3.71339406e+05, 2.02418391e+05, 1.36150922e+05, 8.24756641e+04,\n",
              "        1.96007016e+05, 1.61595500e+05, 1.71401156e+05, 2.00606859e+05,\n",
              "        2.37199281e+05, 1.58720219e+05, 6.36029766e+04, 3.38904938e+05,\n",
              "        5.23524188e+05, 2.86211969e+05, 1.92325922e+05, 3.73764375e+04,\n",
              "        1.31962375e+05, 9.33164844e+04, 1.18443625e+05, 3.00599344e+05,\n",
              "        2.28997641e+05, 3.57954781e+05, 2.36219172e+05, 1.71530031e+05,\n",
              "        2.68483613e+04, 4.79165352e+04, 4.00421953e+04, 5.41032422e+04,\n",
              "        1.07641438e+05, 9.80018828e+04, 7.08349922e+04, 6.55987812e+04,\n",
              "        8.53739609e+04, 1.33836688e+05, 2.72516031e+05, 6.54133000e+05,\n",
              "        4.97211812e+05, 6.11189312e+05, 9.92908312e+05, 9.56691312e+05,\n",
              "        1.09541312e+06, 1.19283188e+06, 3.75002188e+05, 2.22053828e+05,\n",
              "        2.60620156e+05, 1.88108875e+05, 1.87927750e+05, 2.49059219e+05,\n",
              "        3.79191938e+05, 4.07706812e+05, 6.24782625e+05, 9.62080500e+05,\n",
              "        7.95344312e+05, 1.40557766e+05, 2.65226344e+05, 6.26269938e+05,\n",
              "        2.05804641e+05, 6.53085820e+04, 1.33562875e+05, 1.14691586e+05,\n",
              "        2.26913922e+05, 2.61227344e+05, 5.08516812e+05, 4.90744344e+05,\n",
              "        2.38722219e+05, 6.85554844e+04, 7.30579844e+04, 5.36114375e+04,\n",
              "        3.28222750e+05, 6.26447562e+05, 5.17733938e+05, 3.32536906e+05,\n",
              "        1.72656625e+05, 8.28139297e+04, 1.43392859e+05, 8.83170859e+04,\n",
              "        5.15845234e+04, 4.20796602e+04, 1.78049500e+05, 1.17352047e+05,\n",
              "        1.16870359e+05, 1.02416914e+05, 1.67489453e+05, 2.05509344e+05,\n",
              "        2.73222562e+05, 4.73887266e+04, 4.60582461e+04, 5.33724297e+04,\n",
              "        7.87067500e+04, 7.06484219e+04, 5.23133984e+04, 2.50282949e+04,\n",
              "        4.52028984e+04, 4.14731211e+04, 7.00108125e+04, 1.24132516e+05,\n",
              "        1.34167734e+05, 3.38401844e+05, 9.75608906e+04, 1.10023234e+05,\n",
              "        1.99198105e+04, 3.32736992e+04, 2.43713613e+04, 2.76616055e+04,\n",
              "        2.70385918e+04, 1.58114443e+04, 1.16168496e+04, 2.74852637e+04,\n",
              "        2.95268613e+04, 3.95873516e+04, 2.81845449e+04, 3.19157754e+04,\n",
              "        5.60373164e+04, 3.98339648e+04, 5.02201758e+04, 3.19762402e+04,\n",
              "        2.96142949e+04, 4.48334258e+04, 4.32264375e+04, 1.52424111e+04,\n",
              "        2.86075918e+04, 4.23484023e+04, 7.25031172e+04, 1.21924250e+05,\n",
              "        6.41490500e+05, 8.50461375e+05, 1.11135250e+06, 1.22003288e+06,\n",
              "        1.54110100e+06, 2.03451888e+06, 1.76440675e+06, 4.22098719e+05,\n",
              "        1.30513617e+05, 7.30579375e+04, 8.56106875e+04, 7.42399219e+04,\n",
              "        1.49466484e+05, 1.25624111e+04, 2.13100684e+04, 9.81801094e+04,\n",
              "        7.28017891e+04, 1.26064570e+05, 1.01337180e+05, 1.24553477e+05,\n",
              "        7.13406719e+04, 8.41951797e+04, 6.25290625e+04, 8.18796172e+04,\n",
              "        1.27303258e+05, 8.94660078e+04, 1.04100891e+05, 1.66147531e+05,\n",
              "        2.32076047e+05, 4.41966531e+05, 1.75839109e+05, 1.76445078e+05,\n",
              "        9.12171094e+04, 1.67975547e+05, 1.87876500e+05, 1.07636641e+05,\n",
              "        3.26579023e+04, 5.35408945e+04, 1.38144219e+05, 3.28639000e+05,\n",
              "        6.38893000e+05, 1.08279062e+06, 1.55802688e+06, 2.42137000e+06,\n",
              "        2.08980188e+06, 9.83022000e+05, 4.27435062e+05, 5.84014375e+04,\n",
              "        4.02858125e+04, 5.27582695e+04, 1.33694031e+05, 2.41685969e+05,\n",
              "        2.54569906e+05, 2.18475656e+05, 2.39844203e+05, 1.72919391e+05,\n",
              "        9.46134297e+04, 8.14815547e+04, 1.83792781e+05, 3.41631688e+05,\n",
              "        2.67091656e+05, 3.54821562e+05, 1.69607172e+05, 2.63052148e+04,\n",
              "        2.11336660e+04, 1.11685516e+05, 6.08848320e+04, 4.64013945e+04,\n",
              "        6.97663359e+04, 1.16935750e+05, 1.44631531e+05, 1.14627359e+05,\n",
              "        7.92805547e+04, 5.69165547e+04, 3.03999824e+04, 9.47754453e+04,\n",
              "        1.19674555e+05, 2.15399078e+05, 3.15528062e+05, 4.17612281e+05,\n",
              "        3.91753469e+05, 2.83128781e+05, 8.15815625e+04, 5.00838711e+04,\n",
              "        1.01595758e+05, 6.17700664e+04, 3.16725938e+05, 5.76747812e+05,\n",
              "        7.41102938e+05, 1.33218062e+06, 1.56516288e+06, 1.55553262e+06,\n",
              "        1.68517800e+06, 3.21848812e+05, 6.30699648e+04, 7.78368047e+04,\n",
              "        5.12543789e+04, 1.14444672e+05, 1.13520664e+05, 7.03018281e+04,\n",
              "        4.71100312e+04, 4.52434969e+05, 7.79104812e+05, 7.32536688e+05,\n",
              "        1.03909069e+06, 1.56208825e+06, 1.63956662e+06, 1.72790162e+06,\n",
              "        4.00524938e+05, 2.34607969e+05, 1.76724578e+05, 6.78014375e+04,\n",
              "        3.90380906e+05, 9.80138500e+05, 1.03536912e+06, 8.44671688e+05,\n",
              "        1.38894238e+06, 1.98340575e+06, 3.62835000e+06, 2.66920550e+06,\n",
              "        1.42116625e+06, 4.54600656e+05, 3.65966188e+05, 9.46591875e+05,\n",
              "        1.45184338e+06, 1.85899962e+06, 1.89930900e+06, 1.82373088e+06,\n",
              "        1.21286275e+06, 4.94639656e+05, 1.49650112e+06, 2.03091888e+06,\n",
              "        2.25657600e+06, 2.11181391e+05, 1.21395414e+05, 3.04311312e+05,\n",
              "        7.11903125e+05, 6.20310625e+05, 8.05388000e+05, 6.26904125e+05,\n",
              "        5.66389625e+05, 4.27519062e+05, 7.83490438e+05, 3.63024725e+06,\n",
              "        8.18037700e+06, 1.07676110e+07, 1.26563590e+07, 1.33500370e+07,\n",
              "        1.33957210e+07, 6.52510400e+06, 1.60696225e+06, 8.56755125e+05,\n",
              "        1.15616888e+06, 1.69268262e+06, 2.17123725e+06, 3.34217825e+06,\n",
              "        7.84993050e+06, 1.08378820e+07, 1.05727040e+07, 1.01602590e+07,\n",
              "        1.14205950e+07, 1.16179760e+07, 1.73510000e+07, 1.79301540e+07,\n",
              "        2.70626400e+07, 3.72665360e+07, 2.97687480e+07, 2.14580120e+07,\n",
              "        2.44065800e+07, 1.56006630e+07, 1.03771430e+07, 9.93532300e+06,\n",
              "        2.18889420e+07, 3.45002520e+07, 2.72495560e+07, 2.51544575e+06,\n",
              "        2.11410050e+06, 1.65101438e+06, 8.78065100e+06, 1.42263850e+07,\n",
              "        1.50277490e+07, 1.42628800e+07, 1.11521150e+07, 1.31385830e+07,\n",
              "        1.44270310e+07, 2.10808025e+06, 2.37276825e+06, 2.06520375e+06,\n",
              "        1.51271862e+06, 1.63592488e+06, 6.01036300e+06, 1.48181600e+07,\n",
              "        1.51462650e+07, 8.88826800e+06, 1.03656160e+07, 1.24236500e+07,\n",
              "        3.12300380e+07, 4.44121720e+07, 3.35616920e+07, 3.37025520e+07,\n",
              "        4.33024880e+07, 4.04648120e+07, 6.68120760e+07, 2.68143000e+07,\n",
              "        3.37227625e+06, 6.18198100e+06, 3.33603075e+06, 9.76003500e+06,\n",
              "        9.50540900e+06, 1.48300480e+07, 1.08089090e+07, 1.52812330e+07,\n",
              "        1.21298720e+07, 1.47111110e+07, 4.46197040e+07, 6.19226600e+07,\n",
              "        1.05739152e+08, 8.06071200e+07, 3.00356780e+07, 1.30779050e+07,\n",
              "        4.55989850e+06, 1.67929975e+06, 6.99429200e+06, 1.63486880e+07,\n",
              "        1.21498720e+07, 1.21798100e+07, 1.68979380e+07, 2.39387620e+07,\n",
              "        3.73065320e+07, 5.09610680e+07, 4.59522160e+07, 3.20003220e+07,\n",
              "        2.06371660e+07, 8.14506450e+06, 4.14795200e+06, 4.23500700e+06,\n",
              "        3.09785350e+06, 8.37916900e+06, 7.36109850e+06, 7.06096000e+06,\n",
              "        1.57310800e+07, 2.78767000e+07, 4.07340160e+07, 3.28521320e+07,\n",
              "        2.89016620e+07, 3.40229975e+06, 3.84223750e+06, 8.10805100e+06,\n",
              "        1.74778640e+07, 2.13824540e+07, 5.77503650e+06, 7.30699550e+06,\n",
              "        3.70580475e+06, 6.14523812e+05, 1.92766775e+06, 2.29575050e+06,\n",
              "        1.09031838e+06, 1.19422512e+06, 2.01535212e+06, 1.38584200e+06,\n",
              "        2.60915475e+06, 7.48125100e+06, 1.25728590e+07, 1.49188430e+07,\n",
              "        7.64713050e+06, 9.50859600e+06, 1.16537280e+07, 2.15389020e+07,\n",
              "        4.32745680e+07, 6.52255360e+07, 9.30124960e+07, 8.49971200e+07,\n",
              "        5.87007160e+07, 3.43533080e+07, 2.82686400e+07, 9.30443000e+06,\n",
              "        4.66935400e+06, 8.00185850e+06, 2.05975080e+07, 4.07234520e+07,\n",
              "        1.93118560e+07, 4.04351800e+06, 5.92505450e+06, 7.37347950e+06,\n",
              "        2.68488600e+06, 5.27761950e+06, 1.84824825e+06, 4.87236000e+06,\n",
              "        4.87888300e+06, 5.77537050e+06, 1.21702160e+07, 2.70199620e+07,\n",
              "        4.94704520e+07, 6.90301360e+07, 8.50554960e+07], dtype=float32),\n",
              " 'rmse': array([  561.7495 ,   411.29678,   379.64737,   419.52734,   513.34235,\n",
              "          625.7136 ,   632.0809 ,   586.8749 ,   702.77484,   829.9319 ,\n",
              "          958.48987,   894.9798 ,   797.188  ,   803.45245,   598.1142 ,\n",
              "          320.9477 ,   230.89677,   215.83145,   332.18823,   334.13614,\n",
              "          208.65268,   233.53607,   213.50635,   276.5653 ,   200.76714,\n",
              "          164.6441 ,   137.93842,   213.454  ,   206.23889,   257.5365 ,\n",
              "          341.67484,   411.6315 ,   332.39597,   399.6352 ,   402.03345,\n",
              "          351.94003,   388.2749 ,   362.21103,   234.58067,   250.6305 ,\n",
              "          421.2769 ,   163.96413,   159.07056,   160.6378 ,   116.0149 ,\n",
              "           62.30582,    96.06189,   110.63602,   148.617  ,   189.9546 ,\n",
              "          220.89853,   193.55943,   202.62364,   204.30833,   332.40106,\n",
              "          448.06412,   590.0423 ,   598.3822 ,   573.03925,   544.52356,\n",
              "          483.06705,   392.48785,   378.3803 ,   403.22043,   544.47473,\n",
              "          516.60504,   542.39154,   556.8371 ,   250.43527,   293.917  ,\n",
              "          311.37442,   449.23853,   545.02716,   447.9677 ,   356.79877,\n",
              "          364.0196 ,   290.86456,   418.11108,   543.2442 ,   689.8476 ,\n",
              "          849.6431 ,   680.228  ,   364.9037 ,   285.24335,   185.2317 ,\n",
              "          185.79707,   231.70139,   272.47244,   330.6649 ,   528.85016,\n",
              "          468.1019 ,   447.1402 ,   407.3114 ,   285.786  ,   219.63359,\n",
              "          406.5727 ,   426.26086,   575.01733,   680.9015 ,   821.6116 ,\n",
              "          578.71564,   411.51428,   321.27786,   385.1597 ,   546.52124,\n",
              "          609.8789 ,   665.8121 ,   726.5707 ,  1062.757  ,  1003.2593 ,\n",
              "          896.24084,   470.40167,   257.8545 ,   163.14975,   181.57397,\n",
              "          410.6869 ,   419.83063,   547.1423 ,   657.11084,  1383.2822 ,\n",
              "         1975.7382 ,  2508.9888 ,  2343.1655 ,  2501.3538 ,  2507.2266 ,\n",
              "         2690.4958 ,  1442.7177 ,   816.5695 ,   380.31616,   428.26834,\n",
              "          583.9743 ,   549.60754,   609.3762 ,   449.9093 ,   368.98633,\n",
              "          287.18576,   442.7268 ,   401.98944,   414.00623,   447.89157,\n",
              "          487.0311 ,   398.39706,   252.1963 ,   582.1554 ,   723.54974,\n",
              "          534.98785,   438.5498 ,   193.32986,   363.26624,   305.47745,\n",
              "          344.1564 ,   548.2694 ,   478.537  ,   598.2932 ,   486.02383,\n",
              "          414.16183,   163.85469,   218.89845,   200.10545,   232.60104,\n",
              "          328.08755,   313.05252,   266.14844,   256.1226 ,   292.18823,\n",
              "          365.83694,   522.03064,   808.7849 ,   705.13245,   781.78595,\n",
              "          996.4478 ,   978.10596,  1046.6199 ,  1092.1685 ,   612.3742 ,\n",
              "          471.22586,   510.5097 ,   433.7152 ,   433.50635,   499.05832,\n",
              "          615.7856 ,   638.5192 ,   790.4319 ,   980.857  ,   891.82074,\n",
              "          374.91034,   515.0013 ,   791.3722 ,   453.65695,   255.55544,\n",
              "          365.46252,   338.66144,   476.35483,   511.10403,   713.10364,\n",
              "          700.5315 ,   488.59207,   261.83102,   270.2924 ,   231.54144,\n",
              "          572.9073 ,   791.4844 ,   719.5373 ,   576.6601 ,   415.51968,\n",
              "          287.7741 ,   378.6725 ,   297.1819 ,   227.12227,   205.13327,\n",
              "          421.95914,   342.56683,   341.86307,   320.02643,   409.25476,\n",
              "          453.3314 ,   522.707  ,   217.68951,   214.61185,   231.02473,\n",
              "          280.54724,   265.7977 ,   228.72122,   158.20334,   212.60973,\n",
              "          203.6495 ,   264.59555,   352.32443,   366.28912,   581.72314,\n",
              "          312.34738,   331.6975 ,   141.13756,   182.4108 ,   156.1133 ,\n",
              "          166.31778,   164.43416,   125.74356,   107.78149,   165.7868 ,\n",
              "          171.83382,   198.9657 ,   167.88252,   178.64987,   236.72202,\n",
              "          199.58447,   224.09859,   178.81902,   172.08804,   211.73904,\n",
              "          207.90968,   123.46016,   169.13779,   205.78728,   269.26404,\n",
              "          349.1765 ,   800.93097,   922.20465,  1054.207  ,  1104.551  ,\n",
              "         1241.4109 ,  1426.3656 ,  1328.3097 ,   649.6912 ,   361.2667 ,\n",
              "          270.29233,   292.59305,   272.47003,   386.60898,   112.08216,\n",
              "          145.97968,   313.33704,   269.81805,   355.05573,   318.33502,\n",
              "          352.92136,   267.09674,   290.16406,   250.05812,   286.14615,\n",
              "          356.7958 ,   299.10867,   322.6467 ,   407.61197,   481.7427 ,\n",
              "          664.8056 ,   419.33173,   420.05365,   302.0217 ,   409.8482 ,\n",
              "          433.44724,   328.08023,   180.71498,   231.38905,   371.67758,\n",
              "          573.27045,   799.3078 ,  1040.5723 ,  1248.2095 ,  1556.0752 ,\n",
              "         1445.6147 ,   991.4747 ,   653.78516,   241.6639 ,   200.71326,\n",
              "          229.69168,   365.64194,   491.61566,   504.5492 ,   467.4138 ,\n",
              "          489.73892,   415.83575,   307.59296,   285.44974,   428.7106 ,\n",
              "          584.4927 ,   516.8091 ,   595.669  ,   411.83392,   162.18883,\n",
              "          145.37422,   334.19382,   246.74852,   215.40982,   264.13318,\n",
              "          341.95868,   380.30453,   338.56662,   281.56802,   238.5719 ,\n",
              "          174.35591,   307.8562 ,   345.9401 ,   464.11105,   561.7188 ,\n",
              "          646.2293 ,   625.9021 ,   532.09845,   285.62488,   223.79427,\n",
              "          318.74088,   248.53584,   562.78406,   759.43915,   860.87335,\n",
              "         1154.2013 ,  1251.0647 ,  1247.21   ,  1298.144  ,   567.3172 ,\n",
              "          251.13734,   278.99246,   226.3943 ,   338.29672,   336.92828,\n",
              "          265.14493,   217.04845,   672.6329 ,   882.6691 ,   855.88354,\n",
              "         1019.358  ,  1249.8352 ,  1280.4556 ,  1314.4967 ,   632.87036,\n",
              "          484.36346,   420.38623,   260.3871 ,   624.8047 ,   990.0194 ,\n",
              "         1017.5309 ,   919.0602 ,  1178.5339 ,  1408.3344 ,  1904.8229 ,\n",
              "         1633.7703 ,  1192.1268 ,   674.2408 ,   604.95135,   972.9295 ,\n",
              "         1204.9246 ,  1363.4513 ,  1378.1542 ,  1350.4558 ,  1101.3004 ,\n",
              "          703.3062 ,  1223.3156 ,  1425.1031 ,  1502.1904 ,   459.54477,\n",
              "          348.41843,   551.64417,   843.7435 ,   787.598  ,   897.4341 ,\n",
              "          791.77277,   752.5886 ,   653.8494 ,   885.14996,  1905.3208 ,\n",
              "         2860.1357 ,  3281.4038 ,  3557.5776 ,  3653.7703 ,  3660.0164 ,\n",
              "         2554.4282 ,  1267.6602 ,   925.61066,  1075.2529 ,  1301.0314 ,\n",
              "         1473.5118 ,  1828.1625 ,  2801.7727 ,  3292.0938 ,  3251.5696 ,\n",
              "         3187.516  ,  3379.437  ,  3408.5151 ,  4165.453  ,  4234.401  ,\n",
              "         5202.1763 ,  6104.6323 ,  5456.074  ,  4632.2793 ,  4940.3013 ,\n",
              "         3949.7673 ,  3221.3572 ,  3152.0347 ,  4678.562  ,  5873.6914 ,\n",
              "         5220.111  ,  1586.0157 ,  1453.9946 ,  1284.9181 ,  2963.2163 ,\n",
              "         3771.788  ,  3876.564  ,  3776.6228 ,  3339.4783 ,  3624.7183 ,\n",
              "         3798.2932 ,  1451.9229 ,  1540.3793 ,  1437.0817 ,  1229.9263 ,\n",
              "         1279.0327 ,  2451.6042 ,  3849.4363 ,  3891.8203 ,  2981.3198 ,\n",
              "         3219.5676 ,  3524.72   ,  5588.3843 ,  6664.2456 ,  5793.245  ,\n",
              "         5805.3896 ,  6580.4624 ,  6361.196  ,  8173.865  ,  5178.2524 ,\n",
              "         1836.3759 ,  2486.359  ,  1826.4805 ,  3124.1055 ,  3083.0842 ,\n",
              "         3850.98   ,  3287.6904 ,  3909.1218 ,  3482.7966 ,  3835.5066 ,\n",
              "         6679.798  ,  7869.095  , 10282.954  ,  8978.146  ,  5480.4814 ,\n",
              "         3616.3386 ,  2135.3918 ,  1295.8779 ,  2644.6724 ,  4043.351  ,\n",
              "         3485.6667 ,  3489.9585 ,  4110.71   ,  4892.725  ,  6107.907  ,\n",
              "         7138.702  ,  6778.8066 ,  5656.883  ,  4542.8145 ,  2853.9558 ,\n",
              "         2036.6521 ,  2057.9133 ,  1760.072  ,  2894.6794 ,  2713.1345 ,\n",
              "         2657.2466 ,  3966.2424 ,  5279.839  ,  6382.3203 ,  5731.6777 ,\n",
              "         5376.0264 ,  1844.5323 ,  1960.1626 ,  2847.4639 ,  4180.6533 ,\n",
              "         4624.1167 ,  2403.1304 ,  2703.1455 ,  1925.0466 ,   783.91565,\n",
              "         1388.4048 ,  1515.1733 ,  1044.1831 ,  1092.806  ,  1419.631  ,\n",
              "         1177.2179 ,  1615.2878 ,  2735.1875 ,  3545.8228 ,  3862.492  ,\n",
              "         2765.3445 ,  3083.601  ,  3413.7559 ,  4641.002  ,  6578.341  ,\n",
              "         8076.233  ,  9644.299  ,  9219.388  ,  7661.639  ,  5861.1694 ,\n",
              "         5316.826  ,  3050.3164 ,  2160.8687 ,  2828.7556 ,  4538.4478 ,\n",
              "         6381.4927 ,  4394.5254 ,  2010.85   ,  2434.1433 ,  2715.415  ,\n",
              "         1638.5621 ,  2297.307  ,  1359.5029 ,  2207.3423 ,  2208.8193 ,\n",
              "         2403.2    ,  3488.5837 ,  5198.0728 ,  7033.5234 ,  8308.4375 ,\n",
              "         9222.554  ], dtype=float32)}"
            ]
          },
          "metadata": {},
          "execution_count": 74
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Make our evaluation function `evaluate_preds` work for larger horizons"
      ],
      "metadata": {
        "id": "FnWYRvF7TqQH"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model_3_results[\"mae\"].ndim"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "obuLGgoeUOsN",
        "outputId": "6baf1962-1f92-4aae-fa18-d3c9c58982da"
      },
      "execution_count": 75,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "1"
            ]
          },
          "metadata": {},
          "execution_count": 75
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model_2_results[\"mae\"].ndim"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1n5vZk-vUXRr",
        "outputId": "db1ef92e-83d7-4f94-a733-3cc1bc31e1f5"
      },
      "execution_count": 76,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0"
            ]
          },
          "metadata": {},
          "execution_count": 76
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def evaluate_preds(y_true, y_pred):\n",
        "  # Make sure float32 (for metric calculations), if not in float32 tensorflow sometimes gives errors\n",
        "  y_true = tf.cast(y_true, dtype=tf.float32)\n",
        "  y_pred = tf.cast(y_pred, dtype=tf.float32)\n",
        "\n",
        "  # Calculate various metrics\n",
        "  mae = tf.keras.metrics.mean_absolute_error(y_true, y_pred)\n",
        "  mse = tf.keras.metrics.mean_squared_error(y_true, y_pred) # puts and emphasis on outliers (all errors get squared)\n",
        "  rmse = tf.sqrt(mse)\n",
        "  mape = tf.keras.metrics.mean_absolute_percentage_error(y_true, y_pred)\n",
        "  mase = mean_absolute_scaled_error(y_true, y_pred)\n",
        "\n",
        "  # These are the modifications needed:\n",
        "  # Account for different sized metrics (for longer horizons, we want to reduce metrics to a single value)\n",
        "  if mae.ndim > 0:\n",
        "    mae = tf.reduce_mean(mae)\n",
        "    mse = tf.reduce_mean(mse)\n",
        "    rmse = tf.reduce_mean(rmse)\n",
        "    mape = tf.reduce_mean(mape)\n",
        "    mase = tf.reduce_mean(mase)\n",
        "  \n",
        "  return {\"mae\": mae.numpy(),\n",
        "          \"mse\": mse.numpy(),\n",
        "          \"rmse\": rmse.numpy(),\n",
        "          \"mape\": mape.numpy(),\n",
        "          \"mase\": mase.numpy()}"
      ],
      "metadata": {
        "id": "_F5twnBb7QCM"
      },
      "execution_count": 77,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model_3_results = evaluate_preds(y_true = tf.squeeze(test_labels), \n",
        "                                 y_pred = model_3_preds)\n",
        "\n",
        "model_3_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mTYh4-0jbUAV",
        "outputId": "767aeb94-a32c-49d9-9aab-945c04e425cb"
      },
      "execution_count": 78,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'mae': 1237.5065,\n",
              " 'mape': 5.558879,\n",
              " 'mase': 2.2020733,\n",
              " 'mse': 5405199.0,\n",
              " 'rmse': 1425.7478}"
            ]
          },
          "metadata": {},
          "execution_count": 78
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model_3_preds[:5]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pu2UUc1wc-zv",
        "outputId": "7989fb36-0fc7-47d2-bafe-2b14e73d085e"
      },
      "execution_count": 79,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Tensor: shape=(5, 7), dtype=float32, numpy=\n",
              "array([[9004.694 , 9048.1   , 9425.09  , 9258.256 , 9495.797 , 9558.453 ,\n",
              "        9357.3545],\n",
              "       [8735.506 , 8840.303 , 9247.793 , 8885.601 , 9097.186 , 9174.327 ,\n",
              "        9156.818 ],\n",
              "       [8672.508 , 8782.387 , 9123.857 , 8770.371 , 9007.13  , 9003.869 ,\n",
              "        9042.725 ],\n",
              "       [8874.399 , 8784.74  , 9043.902 , 8943.052 , 9033.477 , 9176.489 ,\n",
              "        9039.677 ],\n",
              "       [8825.891 , 8777.44  , 8926.78  , 8870.179 , 9213.232 , 9268.156 ,\n",
              "        8942.487 ]], dtype=float32)>"
            ]
          },
          "metadata": {},
          "execution_count": 79
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "offset = 300\n",
        "plt.figure(figsize=(10,7))\n",
        "plot_time_series(timesteps = X_test[-len(test_windows):], \n",
        "                 values = test_labels[:, 0], \n",
        "                 start = offset,\n",
        "                 format = '-', \n",
        "                 label= \"Test Data\")\n",
        "                 \n",
        "plot_time_series(timesteps = X_test[-len(test_windows):], \n",
        "                 values = model_3_preds, \n",
        "                 start = offset, \n",
        "                 format = '.',\n",
        "                 label= \"model_3_preds\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 442
        },
        "id": "j2wBW4UPbesT",
        "outputId": "429980a1-d875-4c09-f2f7-dfee6444c495"
      },
      "execution_count": 80,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x504 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Below we'll lose some values, as we are using `tf.reduce_mean` to visualise the above plot in a better way, as our model's real predictions are across 7 days"
      ],
      "metadata": {
        "id": "qdg2G5OueCU2"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "offset = 300\n",
        "plt.figure(figsize=(10,7))\n",
        "plot_time_series(timesteps = X_test[-len(test_windows):], \n",
        "                 values = test_labels[:, 0], \n",
        "                 start = offset,\n",
        "                 format = '.', \n",
        "                 label= \"Test Data\")\n",
        "                 \n",
        "plot_time_series(timesteps = X_test[-len(test_windows):], \n",
        "                 values = tf.reduce_mean(model_3_preds, axis=1), \n",
        "                 start = offset, \n",
        "                 format = '-',\n",
        "                 label= \"model_3_preds\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 442
        },
        "id": "hOIN6YKmcuse",
        "outputId": "04a613ee-a3f9-4c3a-8c83-e6a32d7cb592"
      },
      "execution_count": 81,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x504 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Seeing which model has performed the best so far"
      ],
      "metadata": {
        "id": "Ns9JF_-BelSx"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "pd.DataFrame({'naive': naive_results[\"mae\"],\n",
        "              'horizon_1_window_7': model_1_results[\"mae\"],\n",
        "              'horizon_1_window_30': model_2_results[\"mae\"],\n",
        "              'horizon_7_window_30': model_3_results[\"mae\"]}, index =[\"mae\"]).plot(figsize=(10,7), kind = \"bar\") "
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 458
        },
        "id": "Yn20G6f0dmFo",
        "outputId": "1ad46a1c-73ea-4c9c-d7d2-74c47fb23cec"
      },
      "execution_count": 82,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7fc686aefa10>"
            ]
          },
          "metadata": {},
          "execution_count": 82
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x504 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Here, we can see our baseline model (naive model) performs the best, and as we try to predict more into the future the error increases."
      ],
      "metadata": {
        "id": "Snfh1ZlVfzw9"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "If a time series has autocorrelation it means the value at t+1 (the next timestep) is typically close to the value at t (the current timestep). In other words, today's value is probably pretty close to yesterday's value. Of course, this isn't always the case but when it is, a naïve model will often get fairly good results."
      ],
      "metadata": {
        "id": "dZz8sRZiht2x"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Recreating dataset for our LSTM model\n",
        "\n",
        "For the next models we'll use HORIZON = 1 and WINDOW_SIZE = 7"
      ],
      "metadata": {
        "id": "2ISkk9JgGDOe"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "HORIZON = 1 # predict next day of Bitcoin prices\n",
        "WINDOW_SIZE = 7 # use previous week worth of data"
      ],
      "metadata": {
        "id": "wIEcoLGzjy5g"
      },
      "execution_count": 83,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Create windowed data\n",
        "\n",
        "full_windows, full_labels = make_windows(prices, window_size = WINDOW_SIZE, horizon=HORIZON)\n",
        "\n",
        "len(full_windows), len(full_labels)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "AVEljtR0jyx6",
        "outputId": "e934757a-b927-44fe-cada-53236e2032a8"
      },
      "execution_count": 84,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Window step:\n",
            "  [[0 1 2 3 4 5 6 7]]\n",
            "Window indexes:\n",
            " (array([[0, 1, 2, 3, 4, 5, 6, 7],\n",
            "       [1, 2, 3, 4, 5, 6, 7, 8],\n",
            "       [2, 3, 4, 5, 6, 7, 8, 9]]), array([[2777, 2778, 2779, 2780, 2781, 2782, 2783, 2784],\n",
            "       [2778, 2779, 2780, 2781, 2782, 2783, 2784, 2785],\n",
            "       [2779, 2780, 2781, 2782, 2783, 2784, 2785, 2786]]), (2780, 8))\n",
            "Windowed array:\n",
            " [[  123.65499   125.455     108.58483 ...   120.65533   121.795\n",
            "    123.033  ]\n",
            " [  125.455     108.58483   118.67466 ...   121.795     123.033\n",
            "    124.049  ]\n",
            " [  108.58483   118.67466   121.33866 ...   123.033     124.049\n",
            "    125.96116]\n",
            " ...\n",
            " [58788.20968 58102.19143 55715.54665 ... 49764.13208 50032.69314\n",
            "  47885.62525]\n",
            " [58102.19143 55715.54665 56573.55547 ... 50032.69314 47885.62525\n",
            "  45604.61575]\n",
            " [55715.54665 56573.55547 52147.82119 ... 47885.62525 45604.61575\n",
            "  43144.47129]]\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(2780, 2780)"
            ]
          },
          "metadata": {},
          "execution_count": 84
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Create train/test splits\n",
        "\n",
        "train_windows, test_windows, train_labels, test_labels = make_train_test_splits(windows = full_windows, labels = full_labels)\n",
        "\n",
        "len(train_windows), len(test_windows), len(train_labels), len(test_labels)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mfPzjwkTjye6",
        "outputId": "7dd8b254-4ce2-482e-d3fc-984ce144e1fd"
      },
      "execution_count": 85,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(2224, 556, 2224, 556)"
            ]
          },
          "metadata": {},
          "execution_count": 85
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We need and an input shape of: `(batch_size, timesteps, input_dim)`\n",
        "\n",
        "But our data is not in that shape"
      ],
      "metadata": {
        "id": "OYY7hAEIl-ae"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Check data input shape\n",
        "\n",
        "train_windows[0].shape # returns (WINDOW_SIZE, )"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Mlw8NZ-omPhF",
        "outputId": "03d1f3a3-a62b-42f8-d9ed-70177957e24b"
      },
      "execution_count": 86,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(7,)"
            ]
          },
          "metadata": {},
          "execution_count": 86
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "x = tf.constant(train_windows[0]) # Turning our 0th index of train_windows into a tensor\n",
        "x"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZIIJ7IXvmPWR",
        "outputId": "f6de05e2-7f85-4798-b928-d928fd22dfd5"
      },
      "execution_count": 87,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Tensor: shape=(7,), dtype=float64, numpy=\n",
              "array([123.65499, 125.455  , 108.58483, 118.67466, 121.33866, 120.65533,\n",
              "       121.795  ])>"
            ]
          },
          "metadata": {},
          "execution_count": 87
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "expand_dims_layer = layers.Lambda(lambda x: tf.expand_dims(x, axis=1)) # add an extra dimension for `input_dim`"
      ],
      "metadata": {
        "id": "fbW7gGramfw4"
      },
      "execution_count": 88,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Testing out lambda layer \n",
        "\n",
        "print(f\"Original shape: {x.shape}\") # (WINDOW_SIZE)\n",
        "print(f\"Expanded shape: {expand_dims_layer(x).shape}\")\n",
        "print(f\"Original values with expanded shape:\\n {expand_dims_layer(x)}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pLr1uvPcmfjJ",
        "outputId": "7e714ad8-120f-4d3b-d3ed-6801ec37312b"
      },
      "execution_count": 89,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Original shape: (7,)\n",
            "Expanded shape: (7, 1)\n",
            "Original values with expanded shape:\n",
            " [[123.65499]\n",
            " [125.455  ]\n",
            " [108.58483]\n",
            " [118.67466]\n",
            " [121.33866]\n",
            " [120.65533]\n",
            " [121.795  ]]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Model 4: RNN (LSTM)\n",
        "\n",
        "Building RNN model for our time series data"
      ],
      "metadata": {
        "id": "pvLj8q2tiXPR"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "tf.random.set_seed(42)\n",
        "\n",
        "# Building LSTM model with functional API instead of Sequential API as it is a bit more flexible\n",
        "\n",
        "inputs = layers.Input(shape=(WINDOW_SIZE)) # input layer\n",
        "x = layers.Lambda(lambda x: tf.expand_dims(x, axis=1))(inputs) # reshaping the inputs\n",
        "x = layers.LSTM(128, activation =\"relu\")(x) # passing the reshaped inputs through an LSTM\n",
        "output = layers.Dense(HORIZON, activation = \"linear\")(x) # and passing through the output\n",
        "\n",
        "model_4 = tf.keras.Model(inputs = inputs, outputs = output, name = \"model_4_LSTM\")\n",
        "\n",
        "# Compile model\n",
        "\n",
        "model_4.compile(loss=\"mae\",\n",
        "                optimizer = tf.keras.optimizers.Adam())\n",
        "\n",
        "# Fit model\n",
        "\n",
        "model_4.fit(train_windows, \n",
        "            train_labels,\n",
        "            epochs = 100,\n",
        "            verbose = 1,\n",
        "            batch_size = 128,\n",
        "            validation_data = (test_windows, test_labels),\n",
        "            callbacks = [create_model_checkpoint(model_name = model_4.name)]\n",
        "            )\n",
        "\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Zp_UU1-Ufg8a",
        "outputId": "1b662395-15b3-4777-de9f-3c8eb82694b4"
      },
      "execution_count": 90,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "WARNING:tensorflow:Layer lstm will not use cuDNN kernels since it doesn't meet the criteria. It will use a generic GPU kernel as fallback when running on GPU.\n",
            "Epoch 1/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 2534.5144 INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 3s 125ms/step - loss: 2323.6082 - val_loss: 9222.6523\n",
            "Epoch 2/100\n",
            "18/18 [==============================] - ETA: 0s - loss: 695.4739 INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 105ms/step - loss: 695.4739 - val_loss: 1486.0555\n",
            "Epoch 3/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 266.4913INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 103ms/step - loss: 237.3967 - val_loss: 1298.9084\n",
            "Epoch 4/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 192.6114INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 102ms/step - loss: 181.9180 - val_loss: 966.6431\n",
            "Epoch 5/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 173.4022 - val_loss: 970.9295\n",
            "Epoch 6/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 170.0374INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 3s 160ms/step - loss: 173.5414 - val_loss: 941.4746\n",
            "Epoch 7/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 177.6533INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 143ms/step - loss: 170.3566 - val_loss: 898.5215\n",
            "Epoch 8/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 157.6712 - val_loss: 914.8642\n",
            "Epoch 9/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 166.5143 - val_loss: 1079.0776\n",
            "Epoch 10/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 163.3890INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 101ms/step - loss: 162.3369 - val_loss: 877.1690\n",
            "Epoch 11/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 163.0841 - val_loss: 924.1126\n",
            "Epoch 12/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 168.5598 - val_loss: 941.6025\n",
            "Epoch 13/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 166.4726 - val_loss: 934.8436\n",
            "Epoch 14/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 163.9537 - val_loss: 920.2160\n",
            "Epoch 15/100\n",
            "11/18 [=================>............] - ETA: 0s - loss: 151.0251INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 124ms/step - loss: 152.9470 - val_loss: 856.2653\n",
            "Epoch 16/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 147.4420 - val_loss: 898.9914\n",
            "Epoch 17/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 151.0533 - val_loss: 943.1418\n",
            "Epoch 18/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 157.8232 - val_loss: 924.0046\n",
            "Epoch 19/100\n",
            "18/18 [==============================] - 0s 8ms/step - loss: 155.9198 - val_loss: 911.4753\n",
            "Epoch 20/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 155.6456 - val_loss: 862.6680\n",
            "Epoch 21/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 153.1514 - val_loss: 862.5831\n",
            "Epoch 22/100\n",
            "18/18 [==============================] - 0s 8ms/step - loss: 151.4883 - val_loss: 863.8172\n",
            "Epoch 23/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 150.7770 - val_loss: 886.8356\n",
            "Epoch 24/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 148.6539 - val_loss: 861.3271\n",
            "Epoch 25/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 153.1163INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 104ms/step - loss: 146.8467 - val_loss: 845.2782\n",
            "Epoch 26/100\n",
            "11/18 [=================>............] - ETA: 0s - loss: 145.1898INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 105ms/step - loss: 145.9843 - val_loss: 818.6807\n",
            "Epoch 27/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 150.8635INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 102ms/step - loss: 144.4289 - val_loss: 804.3279\n",
            "Epoch 28/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 142.2844INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 123ms/step - loss: 142.4610 - val_loss: 789.0859\n",
            "Epoch 29/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 141.8739 - val_loss: 806.6758\n",
            "Epoch 30/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 142.0521 - val_loss: 794.8836\n",
            "Epoch 31/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 140.2585INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 106ms/step - loss: 138.4690 - val_loss: 773.2812\n",
            "Epoch 32/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 139.7511INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 104ms/step - loss: 137.1986 - val_loss: 772.2625\n",
            "Epoch 33/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 135.9342 - val_loss: 786.4805\n",
            "Epoch 34/100\n",
            "11/18 [=================>............] - ETA: 0s - loss: 142.0283INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 121ms/step - loss: 136.1931 - val_loss: 744.4506\n",
            "Epoch 35/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 133.4513 - val_loss: 756.2970\n",
            "Epoch 36/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 132.2655INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 101ms/step - loss: 134.0992 - val_loss: 735.1729\n",
            "Epoch 37/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 130.6932 - val_loss: 745.9099\n",
            "Epoch 38/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 133.6792INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 104ms/step - loss: 129.2891 - val_loss: 717.2498\n",
            "Epoch 39/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 129.6088 - val_loss: 719.0665\n",
            "Epoch 40/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 129.6111 - val_loss: 732.0632\n",
            "Epoch 41/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 130.4034INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 105ms/step - loss: 127.1393 - val_loss: 702.9354\n",
            "Epoch 42/100\n",
            "14/18 [======================>.......] - ETA: 0s - loss: 123.8576INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 3s 178ms/step - loss: 125.6113 - val_loss: 700.8702\n",
            "Epoch 43/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 125.9959 - val_loss: 704.3682\n",
            "Epoch 44/100\n",
            "11/18 [=================>............] - ETA: 0s - loss: 120.2179INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 105ms/step - loss: 125.2241 - val_loss: 698.7766\n",
            "Epoch 45/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 118.6115INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 103ms/step - loss: 122.5007 - val_loss: 683.0590\n",
            "Epoch 46/100\n",
            "11/18 [=================>............] - ETA: 0s - loss: 130.6197INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 104ms/step - loss: 124.4726 - val_loss: 679.1873\n",
            "Epoch 47/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 122.1862 - val_loss: 705.4241\n",
            "Epoch 48/100\n",
            "11/18 [=================>............] - ETA: 0s - loss: 120.7927INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 122ms/step - loss: 122.0458 - val_loss: 677.0784\n",
            "Epoch 49/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 126.2884INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 105ms/step - loss: 121.1548 - val_loss: 674.6848\n",
            "Epoch 50/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 122.8890 - val_loss: 691.4816\n",
            "Epoch 51/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 122.2994 - val_loss: 737.3929\n",
            "Epoch 52/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 129.3352INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 104ms/step - loss: 121.2944 - val_loss: 657.2365\n",
            "Epoch 53/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 116.8399INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 121ms/step - loss: 119.7668 - val_loss: 652.7970\n",
            "Epoch 54/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 120.1222 - val_loss: 654.4439\n",
            "Epoch 55/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 125.5305INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 103ms/step - loss: 119.2566 - val_loss: 647.2831\n",
            "Epoch 56/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 121.4963INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 102ms/step - loss: 120.0544 - val_loss: 643.7808\n",
            "Epoch 57/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 123.1243 - val_loss: 671.1797\n",
            "Epoch 58/100\n",
            "11/18 [=================>............] - ETA: 0s - loss: 120.3613INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 101ms/step - loss: 118.4059 - val_loss: 640.2656\n",
            "Epoch 59/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 117.7119 - val_loss: 645.3837\n",
            "Epoch 60/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 117.1983INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 122ms/step - loss: 118.5346 - val_loss: 638.7429\n",
            "Epoch 61/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 118.3129 - val_loss: 689.4144\n",
            "Epoch 62/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 121.4350 - val_loss: 748.5565\n",
            "Epoch 63/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 123.1006 - val_loss: 651.8298\n",
            "Epoch 64/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 115.9679 - val_loss: 672.9166\n",
            "Epoch 65/100\n",
            "11/18 [=================>............] - ETA: 0s - loss: 120.7008INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 104ms/step - loss: 118.8341 - val_loss: 631.7538\n",
            "Epoch 66/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 116.2212 - val_loss: 634.5445\n",
            "Epoch 67/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 113.0850INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 102ms/step - loss: 116.1580 - val_loss: 623.8932\n",
            "Epoch 68/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 115.5543 - val_loss: 635.4929\n",
            "Epoch 69/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 120.5866 - val_loss: 626.3387\n",
            "Epoch 70/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 111.8156INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 104ms/step - loss: 114.9188 - val_loss: 619.3181\n",
            "Epoch 71/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 114.8998 - val_loss: 650.9373\n",
            "Epoch 72/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 118.2033 - val_loss: 630.5287\n",
            "Epoch 73/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 115.3612 - val_loss: 623.0369\n",
            "Epoch 74/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 115.0345 - val_loss: 623.0776\n",
            "Epoch 75/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 115.2057 - val_loss: 640.5281\n",
            "Epoch 76/100\n",
            "11/18 [=================>............] - ETA: 0s - loss: 114.5117INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 126ms/step - loss: 114.1402 - val_loss: 614.4062\n",
            "Epoch 77/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 115.5973 - val_loss: 689.8776\n",
            "Epoch 78/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 117.2611 - val_loss: 645.2790\n",
            "Epoch 79/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 112.1790INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 103ms/step - loss: 114.1995 - val_loss: 608.7158\n",
            "Epoch 80/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 112.7125 - val_loss: 614.9222\n",
            "Epoch 81/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 113.3684 - val_loss: 634.8466\n",
            "Epoch 82/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 120.5586INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 103ms/step - loss: 115.9701 - val_loss: 607.8510\n",
            "Epoch 83/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 113.3599 - val_loss: 628.8181\n",
            "Epoch 84/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 114.4113 - val_loss: 628.5579\n",
            "Epoch 85/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 114.8363 - val_loss: 634.7715\n",
            "Epoch 86/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 115.5462 - val_loss: 613.5223\n",
            "Epoch 87/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 115.3621INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 120ms/step - loss: 112.5059 - val_loss: 604.7603\n",
            "Epoch 88/100\n",
            "11/18 [=================>............] - ETA: 0s - loss: 114.8154INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 103ms/step - loss: 111.7706 - val_loss: 601.5748\n",
            "Epoch 89/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 111.8470 - val_loss: 624.0131\n",
            "Epoch 90/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 115.6354 - val_loss: 662.9452\n",
            "Epoch 91/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 116.1836 - val_loss: 636.0310\n",
            "Epoch 92/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 113.1700 - val_loss: 606.0366\n",
            "Epoch 93/100\n",
            "11/18 [=================>............] - ETA: 0s - loss: 113.7486INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 102ms/step - loss: 112.4748 - val_loss: 599.8319\n",
            "Epoch 94/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 111.2637 - val_loss: 601.3256\n",
            "Epoch 95/100\n",
            "10/18 [===============>..............] - ETA: 0s - loss: 108.2715INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 102ms/step - loss: 111.0185 - val_loss: 597.1485\n",
            "Epoch 96/100\n",
            "18/18 [==============================] - 0s 8ms/step - loss: 112.3787 - val_loss: 601.1596\n",
            "Epoch 97/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 111.4700 - val_loss: 626.1573\n",
            "Epoch 98/100\n",
            "18/18 [==============================] - 0s 6ms/step - loss: 111.7326 - val_loss: 601.0142\n",
            "Epoch 99/100\n",
            "12/18 [===================>..........] - ETA: 0s - loss: 109.4926INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/assets\n",
            "WARNING:absl:<keras.layers.recurrent.LSTMCell object at 0x7fc686be0710> has the same name 'LSTMCell' as a built-in Keras object. Consider renaming <class 'keras.layers.recurrent.LSTMCell'> to avoid naming conflicts when loading with `tf.keras.models.load_model`. If renaming is not possible, pass the object in the `custom_objects` parameter of the load function.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r18/18 [==============================] - 2s 126ms/step - loss: 110.7480 - val_loss: 594.6030\n",
            "Epoch 100/100\n",
            "18/18 [==============================] - 0s 7ms/step - loss: 111.9943 - val_loss: 595.9847\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7fc6867c0290>"
            ]
          },
          "metadata": {},
          "execution_count": 90
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Load in best version of model_4 and evaluate on test data\n",
        "model_4 = tf.keras.models.load_model(\"/content/drive/MyDrive/BTC Predictor/model_experiments/model_4_LSTM/\")\n",
        "model_4.evaluate(test_windows, test_labels)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "J2SIb3Poqkqn",
        "outputId": "e4e15109-b4cd-46f2-f8da-3291622421b2"
      },
      "execution_count": 91,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "WARNING:tensorflow:Layer lstm will not use cuDNN kernels since it doesn't meet the criteria. It will use a generic GPU kernel as fallback when running on GPU.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:tensorflow:Layer lstm will not use cuDNN kernels since it doesn't meet the criteria. It will use a generic GPU kernel as fallback when running on GPU.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "18/18 [==============================] - 0s 2ms/step - loss: 594.6030\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "594.6029663085938"
            ]
          },
          "metadata": {},
          "execution_count": 91
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Make prediction with LSTM model \n",
        "model_4_preds = make_preds(model_4, test_windows)\n",
        "model_4_preds[:5]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mIWkDWJGz6vt",
        "outputId": "bbc4e852-aefe-4e06-8845-fa600e37d0e3"
      },
      "execution_count": 92,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Tensor: shape=(5,), dtype=float32, numpy=array([8978.46 , 8810.951, 8997.779, 8836.308, 8731.047], dtype=float32)>"
            ]
          },
          "metadata": {},
          "execution_count": 92
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Evaluate model 4 predictions\n",
        "\n",
        "model_4_results = evaluate_preds(y_true = tf.squeeze(test_labels), \n",
        "                               y_pred = model_4_preds)\n",
        "\n",
        "model_4_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hJ2pvIZB0MJd",
        "outputId": "201468a6-f79d-433b-83b7-9a20212ebd4b"
      },
      "execution_count": 93,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'mae': 594.60297,\n",
              " 'mape': 2.6717565,\n",
              " 'mase': 1.0445527,\n",
              " 'mse': 1267231.9,\n",
              " 'rmse': 1125.7139}"
            ]
          },
          "metadata": {},
          "execution_count": 93
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model_1_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GkuiVuy70fD2",
        "outputId": "f2a1813b-7f43-409c-f2aa-89c7496ebbca"
      },
      "execution_count": 94,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'mae': 568.95105,\n",
              " 'mape': 2.5448983,\n",
              " 'mase': 0.9994894,\n",
              " 'mse': 1171743.8,\n",
              " 'rmse': 1082.4711}"
            ]
          },
          "metadata": {},
          "execution_count": 94
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Preparing Data for Model 5 (Ensemble) using `tf.data` API\n",
        "\n",
        "Use: \n",
        "\n",
        "To make our data performant, i.e. to make our input data load as fast as possible"
      ],
      "metadata": {
        "id": "jh5hBqRi3D3k"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "HORIZON = 1\n",
        "WINDOW_SIZE = 7"
      ],
      "metadata": {
        "id": "Lz5LKF8r28CH"
      },
      "execution_count": 95,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Create data inputs\n",
        "\n",
        "bitcoin_prices.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 237
        },
        "id": "ZuGmtMk03wdz",
        "outputId": "ccd07655-76c8-4f8d-912e-6a324e699e56"
      },
      "execution_count": 96,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                Price\n",
              "Date                 \n",
              "2013-10-01  123.65499\n",
              "2013-10-02  125.45500\n",
              "2013-10-03  108.58483\n",
              "2013-10-04  118.67466\n",
              "2013-10-05  121.33866"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-f8dcc050-7c0e-4f64-b709-2e8f71b5fc89\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Price</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Date</th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2013-10-01</th>\n",
              "      <td>123.65499</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2013-10-02</th>\n",
              "      <td>125.45500</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2013-10-03</th>\n",
              "      <td>108.58483</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2013-10-04</th>\n",
              "      <td>118.67466</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2013-10-05</th>\n",
              "      <td>121.33866</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-f8dcc050-7c0e-4f64-b709-2e8f71b5fc89')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-f8dcc050-7c0e-4f64-b709-2e8f71b5fc89 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-f8dcc050-7c0e-4f64-b709-2e8f71b5fc89');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 96
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Add windowed columns\n",
        "\n",
        "bitcoin_prices_ensemble = bitcoin_prices.copy()\n",
        "for i in range(WINDOW_SIZE):\n",
        "  bitcoin_prices_ensemble[f\"Price+{i+1}\"] = bitcoin_prices_ensemble[\"Price\"].shift(periods=i+1)\n",
        "\n",
        "bitcoin_prices_ensemble.head() "
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 237
        },
        "id": "BG4ABJnZ3wae",
        "outputId": "fbea94f2-a121-4915-f9f8-fb88d7f4e1a6"
      },
      "execution_count": 97,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                Price    Price+1    Price+2    Price+3    Price+4  Price+5  \\\n",
              "Date                                                                         \n",
              "2013-10-01  123.65499        NaN        NaN        NaN        NaN      NaN   \n",
              "2013-10-02  125.45500  123.65499        NaN        NaN        NaN      NaN   \n",
              "2013-10-03  108.58483  125.45500  123.65499        NaN        NaN      NaN   \n",
              "2013-10-04  118.67466  108.58483  125.45500  123.65499        NaN      NaN   \n",
              "2013-10-05  121.33866  118.67466  108.58483  125.45500  123.65499      NaN   \n",
              "\n",
              "            Price+6  Price+7  \n",
              "Date                          \n",
              "2013-10-01      NaN      NaN  \n",
              "2013-10-02      NaN      NaN  \n",
              "2013-10-03      NaN      NaN  \n",
              "2013-10-04      NaN      NaN  \n",
              "2013-10-05      NaN      NaN  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-89c8ae74-427b-40a8-9535-a3df38f29f6b\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Price</th>\n",
              "      <th>Price+1</th>\n",
              "      <th>Price+2</th>\n",
              "      <th>Price+3</th>\n",
              "      <th>Price+4</th>\n",
              "      <th>Price+5</th>\n",
              "      <th>Price+6</th>\n",
              "      <th>Price+7</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Date</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2013-10-01</th>\n",
              "      <td>123.65499</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2013-10-02</th>\n",
              "      <td>125.45500</td>\n",
              "      <td>123.65499</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2013-10-03</th>\n",
              "      <td>108.58483</td>\n",
              "      <td>125.45500</td>\n",
              "      <td>123.65499</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2013-10-04</th>\n",
              "      <td>118.67466</td>\n",
              "      <td>108.58483</td>\n",
              "      <td>125.45500</td>\n",
              "      <td>123.65499</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2013-10-05</th>\n",
              "      <td>121.33866</td>\n",
              "      <td>118.67466</td>\n",
              "      <td>108.58483</td>\n",
              "      <td>125.45500</td>\n",
              "      <td>123.65499</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "      <td>NaN</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-89c8ae74-427b-40a8-9535-a3df38f29f6b')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-89c8ae74-427b-40a8-9535-a3df38f29f6b button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-89c8ae74-427b-40a8-9535-a3df38f29f6b');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 97
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Make features and labels\n",
        "\n",
        "X = bitcoin_prices_ensemble.dropna().drop(\"Price\", axis = 1)\n",
        "y = bitcoin_prices_ensemble.dropna()[\"Price\"]\n",
        "\n",
        "# Make train and test splits\n",
        "\n",
        "split_size = int(len(X) *  0.8)\n",
        "X_train, y_train = X[:split_size], y[:split_size]\n",
        "X_test, y_test = X[split_size:], y[split_size:]\n",
        "\n",
        "len(X_train), len(y_train), len(X_test), len(y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5AQvTnXy3wXR",
        "outputId": "03f38d39-4ff6-4414-9451-c12c0a67bcfd"
      },
      "execution_count": 98,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(2224, 2224, 556, 556)"
            ]
          },
          "metadata": {},
          "execution_count": 98
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Making our dataset performant using tf.data api\n",
        "\n",
        "train_features_dataset = tf.data.Dataset.from_tensor_slices(X_train) # from_tensor_slices() creates a dataset whose elements \n",
        "                                                                     # are slices of the given tensors\n",
        "train_labels_dataset = tf.data.Dataset.from_tensor_slices(y_train)\n",
        "\n",
        "test_features_dataset = tf.data.Dataset.from_tensor_slices(X_test)\n",
        "test_labels_dataset = tf.data.Dataset.from_tensor_slices(y_test)\n",
        "\n",
        "# Compining labels and features by zipping together into a tuple -> (features, labels)\n",
        "\n",
        "train_dataset = tf.data.Dataset.zip((train_features_dataset, train_labels_dataset)) # zip() creates a Dataset by zipping together the given datasets\n",
        "test_dataset = tf.data.Dataset.zip((test_features_dataset, test_labels_dataset))\n",
        "\n",
        "# Batch and prefetch\n",
        "\n",
        "BATCH_SIZE = 1024\n",
        "\n",
        "# Batch -  The tf.data.Dataset.batch() function is used to group the elements into batches.\n",
        "# Prefetch - Creates a Dataset that prefetches elements from this dataset. \n",
        "#            Most dataset input pipelines should end with a call to prefetch. This allows later elements to be prepared while the current element is being processed. This often improves latency and throughput, at the cost of using additional memory to store prefetched elements.\n",
        "train_dataset = train_dataset.batch(BATCH_SIZE).prefetch(tf.data.AUTOTUNE)\n",
        "test_dataset = test_dataset.batch(BATCH_SIZE).prefetch(tf.data.AUTOTUNE)\n",
        "\n",
        "train_dataset, test_dataset"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "h_R4k3G43wLt",
        "outputId": "f237d30f-b667-4863-e90e-dd5386a226fa"
      },
      "execution_count": 99,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(<PrefetchDataset element_spec=(TensorSpec(shape=(None, 7), dtype=tf.float64, name=None), TensorSpec(shape=(None,), dtype=tf.float64, name=None))>,\n",
              " <PrefetchDataset element_spec=(TensorSpec(shape=(None, 7), dtype=tf.float64, name=None), TensorSpec(shape=(None,), dtype=tf.float64, name=None))>)"
            ]
          },
          "metadata": {},
          "execution_count": 99
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "1. batch( batch_size, name=None)\n",
        "\n",
        "  - Combines consecutive elements of this dataset into batches. \n",
        "\n",
        "2. prefetch( buffer_size, name=None)\n",
        "\n",
        "  - Creates a Dataset that prefetches elements from this dataset.\n",
        "  - Most dataset input pipelines should end with a call to prefetch. This allows later elements to be prepared while the current element is being processed. This often improves latency and throughput, at the cost of using additional memory to store prefetched elements."
      ],
      "metadata": {
        "id": "Tb90cd0tEvUX"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Model 5: Creating an ensemble (stacking different models together)\n",
        "\n",
        "An ensemble leverages the wisdom of the crowd effect\n"
      ],
      "metadata": {
        "id": "1hVitgaf0-dr"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Constructing and fitting an ensemble of models (using different loss functions)"
      ],
      "metadata": {
        "id": "mRAR5DlO2FHJ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def get_ensemble_models(horizon=HORIZON, \n",
        "                        train_data=train_dataset,\n",
        "                        test_data=test_dataset,\n",
        "                        num_iter=10, # number of models we are going to train\n",
        "                        num_epochs=1000, \n",
        "                        loss_fns=[\"mae\", \"mse\", \"mape\"]):\n",
        "  \"\"\"\n",
        "  Returns a list of num_iter models each trained on MAE, MSE and MAPE loss.\n",
        "\n",
        "  For example, if num_iter=10, a list of 30 trained models will be returned:\n",
        "  10 * len([\"mae\", \"mse\", \"mape\"]).\n",
        "  \"\"\"\n",
        "  # Make empty list for trained ensemble models\n",
        "  ensemble_models = []\n",
        "\n",
        "  # Create num_iter number of models per loss function\n",
        "  for i in range(num_iter):\n",
        "    # Build and fit a new model with a different loss function\n",
        "    for loss_function in loss_fns:\n",
        "      print(f\"Optimizing model by reducing: {loss_function} for {num_epochs} epochs, model number: {i}\")\n",
        "\n",
        "      # Construct a simple model (similar to model_1)\n",
        "      model = tf.keras.Sequential([\n",
        "        # Initialize layers with normal (Gaussian) distribution so we can use the models for prediction\n",
        "        layers.Dense(128, kernel_initializer=\"he_normal\", activation=\"relu\"), \n",
        "        layers.Dense(128, kernel_initializer=\"he_normal\", activation=\"relu\"),\n",
        "        layers.Dense(HORIZON)                                 \n",
        "      ])\n",
        "\n",
        "      # Compile simple model with current loss function\n",
        "      model.compile(loss=loss_function,\n",
        "                    optimizer=tf.keras.optimizers.Adam(),\n",
        "                    metrics=[\"mae\", \"mse\"])\n",
        "      \n",
        "      # Fit model\n",
        "      model.fit(train_data,\n",
        "                epochs=num_epochs,\n",
        "                verbose=0,\n",
        "                validation_data=test_data,\n",
        "                # Add callbacks to prevent training from going/stalling for too long\n",
        "                callbacks=[tf.keras.callbacks.EarlyStopping(monitor=\"val_loss\",\n",
        "                                                            patience=200,\n",
        "                                                            restore_best_weights=True),\n",
        "                # The above Early Stopping callback means: if \"val_loss\" doesn't improve for 200 epochs, stop training and\n",
        "                # restore our model to the best_weights, so that we are not burning through extra epochs that we don't have to\n",
        "                           tf.keras.callbacks.ReduceLROnPlateau(monitor=\"val_loss\",\n",
        "                                                                patience=100,\n",
        "                                                                verbose=1)])\n",
        "                # The above ReduceLROnPlateau means: reduces the learning rate by a factor of 0.1, if \"val_loss\" doesn't \n",
        "                # improve after a 100 epochs\n",
        "      \n",
        "      # Append fitted model to list of ensemble models\n",
        "      ensemble_models.append(model)\n",
        "\n",
        "  return ensemble_models # return list of trained models"
      ],
      "metadata": {
        "id": "ZsxUC12qTTUS"
      },
      "execution_count": 100,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "%%time\n",
        "# Get list of trained ensemble models\n",
        "ensemble_models = get_ensemble_models(num_iter = 5,\n",
        "                                      num_epochs = 1000)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "j9kfKSgdNJYl",
        "outputId": "d74164a7-26ba-46ea-df07-17e0db36e32d"
      },
      "execution_count": 101,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Optimizing model by reducing: mae for 1000 epochs, model number: 0\n",
            "\n",
            "Epoch 650: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n",
            "\n",
            "Epoch 791: ReduceLROnPlateau reducing learning rate to 1.0000000474974514e-05.\n",
            "\n",
            "Epoch 891: ReduceLROnPlateau reducing learning rate to 1.0000000656873453e-06.\n",
            "Optimizing model by reducing: mse for 1000 epochs, model number: 0\n",
            "\n",
            "Epoch 624: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n",
            "\n",
            "Epoch 724: ReduceLROnPlateau reducing learning rate to 1.0000000474974514e-05.\n",
            "Optimizing model by reducing: mape for 1000 epochs, model number: 0\n",
            "\n",
            "Epoch 361: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n",
            "\n",
            "Epoch 715: ReduceLROnPlateau reducing learning rate to 1.0000000474974514e-05.\n",
            "\n",
            "Epoch 815: ReduceLROnPlateau reducing learning rate to 1.0000000656873453e-06.\n",
            "Optimizing model by reducing: mae for 1000 epochs, model number: 1\n",
            "\n",
            "Epoch 563: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n",
            "Optimizing model by reducing: mse for 1000 epochs, model number: 1\n",
            "\n",
            "Epoch 393: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n",
            "\n",
            "Epoch 493: ReduceLROnPlateau reducing learning rate to 1.0000000474974514e-05.\n",
            "Optimizing model by reducing: mape for 1000 epochs, model number: 1\n",
            "\n",
            "Epoch 267: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n",
            "\n",
            "Epoch 367: ReduceLROnPlateau reducing learning rate to 1.0000000474974514e-05.\n",
            "Optimizing model by reducing: mae for 1000 epochs, model number: 2\n",
            "\n",
            "Epoch 983: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n",
            "Optimizing model by reducing: mse for 1000 epochs, model number: 2\n",
            "\n",
            "Epoch 150: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n",
            "\n",
            "Epoch 250: ReduceLROnPlateau reducing learning rate to 1.0000000474974514e-05.\n",
            "Optimizing model by reducing: mape for 1000 epochs, model number: 2\n",
            "\n",
            "Epoch 132: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n",
            "\n",
            "Epoch 594: ReduceLROnPlateau reducing learning rate to 1.0000000474974514e-05.\n",
            "\n",
            "Epoch 700: ReduceLROnPlateau reducing learning rate to 1.0000000656873453e-06.\n",
            "\n",
            "Epoch 800: ReduceLROnPlateau reducing learning rate to 1.0000001111620805e-07.\n",
            "Optimizing model by reducing: mae for 1000 epochs, model number: 3\n",
            "Optimizing model by reducing: mse for 1000 epochs, model number: 3\n",
            "\n",
            "Epoch 284: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n",
            "\n",
            "Epoch 384: ReduceLROnPlateau reducing learning rate to 1.0000000474974514e-05.\n",
            "Optimizing model by reducing: mape for 1000 epochs, model number: 3\n",
            "\n",
            "Epoch 265: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n",
            "\n",
            "Epoch 365: ReduceLROnPlateau reducing learning rate to 1.0000000474974514e-05.\n",
            "Optimizing model by reducing: mae for 1000 epochs, model number: 4\n",
            "Optimizing model by reducing: mse for 1000 epochs, model number: 4\n",
            "\n",
            "Epoch 699: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n",
            "\n",
            "Epoch 799: ReduceLROnPlateau reducing learning rate to 1.0000000474974514e-05.\n",
            "Optimizing model by reducing: mape for 1000 epochs, model number: 4\n",
            "\n",
            "Epoch 158: ReduceLROnPlateau reducing learning rate to 0.00010000000474974513.\n",
            "\n",
            "Epoch 632: ReduceLROnPlateau reducing learning rate to 1.0000000474974514e-05.\n",
            "\n",
            "Epoch 800: ReduceLROnPlateau reducing learning rate to 1.0000000656873453e-06.\n",
            "\n",
            "Epoch 900: ReduceLROnPlateau reducing learning rate to 1.0000001111620805e-07.\n",
            "CPU times: user 5min 33s, sys: 25.2 s, total: 5min 58s\n",
            "Wall time: 5min 29s\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!nvidia-smi"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VGmEZheiOyuU",
        "outputId": "900dbc8d-0aaa-4e7a-93a9-225f9d95e2c4"
      },
      "execution_count": 102,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mon Aug  8 20:37:09 2022       \n",
            "+-----------------------------------------------------------------------------+\n",
            "| NVIDIA-SMI 460.32.03    Driver Version: 460.32.03    CUDA Version: 11.2     |\n",
            "|-------------------------------+----------------------+----------------------+\n",
            "| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\n",
            "| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |\n",
            "|                               |                      |               MIG M. |\n",
            "|===============================+======================+======================|\n",
            "|   0  Tesla T4            Off  | 00000000:00:04.0 Off |                    0 |\n",
            "| N/A   46C    P0    25W /  70W |    462MiB / 15109MiB |      7%      Default |\n",
            "|                               |                      |                  N/A |\n",
            "+-------------------------------+----------------------+----------------------+\n",
            "                                                                               \n",
            "+-----------------------------------------------------------------------------+\n",
            "| Processes:                                                                  |\n",
            "|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |\n",
            "|        ID   ID                                                   Usage      |\n",
            "|=============================================================================|\n",
            "+-----------------------------------------------------------------------------+\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "ensemble_models"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "R-Ztgdt-RD-U",
        "outputId": "bacd3e2a-a6dd-4b59-864c-a7e398c3e916"
      },
      "execution_count": 103,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<keras.engine.sequential.Sequential at 0x7fc69e2b2210>,\n",
              " <keras.engine.sequential.Sequential at 0x7fc678389850>,\n",
              " <keras.engine.sequential.Sequential at 0x7fc6865f2bd0>,\n",
              " <keras.engine.sequential.Sequential at 0x7fc6780c0e90>,\n",
              " <keras.engine.sequential.Sequential at 0x7fc67850b650>,\n",
              " <keras.engine.sequential.Sequential at 0x7fc688081290>,\n",
              " <keras.engine.sequential.Sequential at 0x7fc6866afe90>,\n",
              " <keras.engine.sequential.Sequential at 0x7fc674669f50>,\n",
              " <keras.engine.sequential.Sequential at 0x7fc66fdebf10>,\n",
              " <keras.engine.sequential.Sequential at 0x7fc67056bd90>,\n",
              " <keras.engine.sequential.Sequential at 0x7fc670478850>,\n",
              " <keras.engine.sequential.Sequential at 0x7fc686544f90>,\n",
              " <keras.engine.sequential.Sequential at 0x7fc6747c4690>,\n",
              " <keras.engine.sequential.Sequential at 0x7fc67843d250>,\n",
              " <keras.engine.sequential.Sequential at 0x7fc67024d050>]"
            ]
          },
          "metadata": {},
          "execution_count": 103
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Make predictions with our ensemble model"
      ],
      "metadata": {
        "id": "37w7k1sYQg_A"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Create a function which uses a list of trained models to make and return a list of predictions\n",
        "def make_ensemble_preds(ensemble_models, data):\n",
        "  ensemble_preds = []\n",
        "  for model in ensemble_models:\n",
        "    preds = model.predict(data)\n",
        "    ensemble_preds.append(preds)\n",
        "  return tf.constant(tf.squeeze(ensemble_preds))\n"
      ],
      "metadata": {
        "id": "raz2IUDYQb2P"
      },
      "execution_count": 104,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Create a list of ensemble preds\n",
        "%%time\n",
        "\n",
        "ensemble_preds = make_ensemble_preds(ensemble_models = ensemble_models, \n",
        "                                     data = test_dataset)\n",
        "\n",
        "ensemble_preds"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GcloFzq_SNQ1",
        "outputId": "a1e9a19f-ee65-4e50-c4e8-9bc59b4effa7"
      },
      "execution_count": 105,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "WARNING:tensorflow:5 out of the last 22 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fc6780c98c0> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:tensorflow:5 out of the last 22 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fc6780c98c0> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "WARNING:tensorflow:6 out of the last 23 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fc687db1830> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:tensorflow:6 out of the last 23 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fc687db1830> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "CPU times: user 610 ms, sys: 37.2 ms, total: 647 ms\n",
            "Wall time: 637 ms\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "ensemble_preds"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Kcn9cqSXSU64",
        "outputId": "580e5fc5-22d8-48d6-cd50-4c616904712b"
      },
      "execution_count": 106,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Tensor: shape=(15, 556), dtype=float32, numpy=\n",
              "array([[ 8778.674 ,  8843.493 ,  9064.171 , ..., 49279.664 , 48609.297 ,\n",
              "        46203.156 ],\n",
              "       [ 8736.658 ,  8787.107 ,  9039.8125, ..., 50116.11  , 48259.85  ,\n",
              "        45876.703 ],\n",
              "       [ 8780.267 ,  8805.378 ,  9092.468 , ..., 50267.844 , 48242.688 ,\n",
              "        45886.45  ],\n",
              "       ...,\n",
              "       [ 8661.287 ,  8642.298 ,  9134.07  , ..., 49868.92  , 48643.95  ,\n",
              "        45533.344 ],\n",
              "       [ 8812.776 ,  8754.445 ,  9002.784 , ..., 49482.824 , 48219.027 ,\n",
              "        46167.91  ],\n",
              "       [ 8809.392 ,  8810.626 ,  9067.475 , ..., 50012.508 , 47947.445 ,\n",
              "        45541.402 ]], dtype=float32)>"
            ]
          },
          "metadata": {},
          "execution_count": 106
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now we've got a set of ensemble predictions, we can evaluate them against the ground truth values.\n",
        "\n",
        "However, since we've trained 15 models, there's going to be 15 sets of predictions. Rather than comparing every set of predictions to the ground truth, let's take the median (you could also take the mean too but the median is usually more robust than the mean)."
      ],
      "metadata": {
        "id": "GkfngkONXelx"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "ensemble_median = np.median(ensemble_preds, axis=0)\n",
        "ensemble_median[:10]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GaoXbWi2VBBr",
        "outputId": "4d8069f7-f829-4cb6-85f4-3ad91e19bfb4"
      },
      "execution_count": 107,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([8794.825, 8775.06 , 9057.606, 8712.127, 8741.391, 8716.866,\n",
              "       8647.917, 8483.576, 8452.958, 8520.001], dtype=float32)"
            ]
          },
          "metadata": {},
          "execution_count": 107
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Evaluate ensemble model(s) predictions\n",
        "ensemble_results = evaluate_preds(y_true = y_test, \n",
        "                                  y_pred = ensemble_median)\n",
        "\n",
        "ensemble_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ym-MN2_WVlei",
        "outputId": "3f5e1a46-2e7b-4b7e-84f1-7ceb8a851ce4"
      },
      "execution_count": 108,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'mae': 565.0479,\n",
              " 'mape': 2.554046,\n",
              " 'mase': 0.9926327,\n",
              " 'mse': 1131961.5,\n",
              " 'rmse': 1063.9368}"
            ]
          },
          "metadata": {},
          "execution_count": 108
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "It looks like the ensemble model is the best performing model so far"
      ],
      "metadata": {
        "id": "FYZ8XpvpYQfs"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Why forecasting doesn't work (the turkey problem)\n",
        "\n",
        "You could have the historical data of Bitcoin for its entire existence and build a model which predicts it perfectly.\n",
        "\n",
        "But then one day for some unknown and unpredictable reason, the price of Bitcoin plummets 100x in a single day.\n",
        "\n",
        "Just like a turkey could live 1000 days eating healthy, and on the day of thanksgiving (1001st day) it may become a roasted turkey."
      ],
      "metadata": {
        "id": "eOZ7WFzGXymM"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "btc_price_turkeys = bitcoin_prices.copy()"
      ],
      "metadata": {
        "id": "-135cOyxYPBA"
      },
      "execution_count": 109,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "len(btc_price_turkeys)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uPpPcUfhZGlP",
        "outputId": "b6b42f1f-d2c9-4637-de08-6c12b5969668"
      },
      "execution_count": 110,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2787"
            ]
          },
          "metadata": {},
          "execution_count": 110
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "btc_price_turkey = btc_price_turkeys[\"Price\"].values.tolist()\n",
        "\n",
        "\n",
        "# Introducing the above problem in our BTC data (price of BTC falls 100x in 1 day)\n",
        "\n",
        "btc_price_turkey[-1] = btc_price_turkey[-1] / 100"
      ],
      "metadata": {
        "id": "MCqpvaTmecrT"
      },
      "execution_count": 111,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "btc_price_turkey[-10:]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0xAVUE9efTsp",
        "outputId": "1a173764-2102-4f09-ad3f-1dd6505e9f64"
      },
      "execution_count": 112,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[58788.20968,\n",
              " 58102.19143,\n",
              " 55715.54665,\n",
              " 56573.55547,\n",
              " 52147.82119,\n",
              " 49764.13208,\n",
              " 50032.69314,\n",
              " 47885.62525,\n",
              " 45604.61575,\n",
              " 431.4447129]"
            ]
          },
          "metadata": {},
          "execution_count": 112
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Get the timesteps for the turkey problem\n",
        "\n",
        "btc_timesteps_turkey = np.array(bitcoin_prices.index)\n",
        "btc_timesteps_turkey[-10:]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XnOp550N0Mrx",
        "outputId": "4fea879b-212b-43c0-8586-0e5605d0d931"
      },
      "execution_count": 113,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array(['2021-05-09T00:00:00.000000000', '2021-05-10T00:00:00.000000000',\n",
              "       '2021-05-11T00:00:00.000000000', '2021-05-12T00:00:00.000000000',\n",
              "       '2021-05-13T00:00:00.000000000', '2021-05-14T00:00:00.000000000',\n",
              "       '2021-05-15T00:00:00.000000000', '2021-05-16T00:00:00.000000000',\n",
              "       '2021-05-17T00:00:00.000000000', '2021-05-18T00:00:00.000000000'],\n",
              "      dtype='datetime64[ns]')"
            ]
          },
          "metadata": {},
          "execution_count": 113
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Plotting our turkey data\n",
        "\n",
        "plt.figure(figsize=(10, 7))\n",
        "plot_time_series(timesteps = btc_timesteps_turkey, values=btc_price_turkey, format='-', start = 2500, label = \"BTC Price + Turkey Problem\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 442
        },
        "id": "UFw0jAmp0zz9",
        "outputId": "682fe455-4460-4a7a-9b15-e6953c240bd9"
      },
      "execution_count": 114,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x504 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Preparing data for the turkey model"
      ],
      "metadata": {
        "id": "j4mIYbpzHFRq"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "HORIZON = 1\n",
        "WINDOW_SIZE = 7"
      ],
      "metadata": {
        "id": "8tJYoHhi2KXJ"
      },
      "execution_count": 115,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "full_windows, full_labels = make_windows(np.array(btc_price_turkey), window_size = WINDOW_SIZE, horizon = HORIZON)\n",
        "\n",
        "len(full_windows), len(full_labels)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ym_NRAr41XtE",
        "outputId": "c49ac681-1202-4fcd-9550-ad77a6741eb2"
      },
      "execution_count": 116,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Window step:\n",
            "  [[0 1 2 3 4 5 6 7]]\n",
            "Window indexes:\n",
            " (array([[0, 1, 2, 3, 4, 5, 6, 7],\n",
            "       [1, 2, 3, 4, 5, 6, 7, 8],\n",
            "       [2, 3, 4, 5, 6, 7, 8, 9]]), array([[2777, 2778, 2779, 2780, 2781, 2782, 2783, 2784],\n",
            "       [2778, 2779, 2780, 2781, 2782, 2783, 2784, 2785],\n",
            "       [2779, 2780, 2781, 2782, 2783, 2784, 2785, 2786]]), (2780, 8))\n",
            "Windowed array:\n",
            " [[  123.65499     125.455       108.58483   ...   120.65533\n",
            "    121.795       123.033    ]\n",
            " [  125.455       108.58483     118.67466   ...   121.795\n",
            "    123.033       124.049    ]\n",
            " [  108.58483     118.67466     121.33866   ...   123.033\n",
            "    124.049       125.96116  ]\n",
            " ...\n",
            " [58788.20968   58102.19143   55715.54665   ... 49764.13208\n",
            "  50032.69314   47885.62525  ]\n",
            " [58102.19143   55715.54665   56573.55547   ... 50032.69314\n",
            "  47885.62525   45604.61575  ]\n",
            " [55715.54665   56573.55547   52147.82119   ... 47885.62525\n",
            "  45604.61575     431.4447129]]\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(2780, 2780)"
            ]
          },
          "metadata": {},
          "execution_count": 116
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Create train and test sets for turkey problem data\n",
        "\n",
        "X_train, X_test, y_train, y_test = make_train_test_splits(windows = full_windows, labels = full_labels)\n",
        "\n",
        "len(X_train), len(X_test), len(y_train), len(y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VAbXTre311Of",
        "outputId": "03b63b01-dee3-40c5-90aa-f651317482ff"
      },
      "execution_count": 117,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(2224, 556, 2224, 556)"
            ]
          },
          "metadata": {},
          "execution_count": 117
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Model 6: Turkey Model"
      ],
      "metadata": {
        "id": "JPl8bqn_HQ8T"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Clone model 1 architecture for turkey model and fit the turkey model on the turkey data\n",
        "\n",
        "turkey_model = tf.keras.models.clone_model(model_1)\n",
        "turkey_model._name = \"model_6_turkey_model\" # To access and change the name feature in model.keras.Sequential() in model 1\n",
        "\n",
        "# Compile the model\n",
        "turkey_model.compile(loss=\"mae\",\n",
        "                     optimizer = tf.keras.optimizers.Adam())\n",
        "\n",
        "# Fit the model\n",
        "\n",
        "turkey_model.fit(x = X_train,\n",
        "                 y = y_train,\n",
        "                 batch_size = 128,\n",
        "                 epochs = 100,\n",
        "                 verbose = 0,\n",
        "                 validation_data = [X_test, y_test],\n",
        "                 callbacks = [create_model_checkpoint(model_name = \"model_6_turkey_model\")])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "SLAxhEkv2l8y",
        "outputId": "f570885a-4cfc-4fc5-a329-f5f592814773"
      },
      "execution_count": 118,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/assets\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7fc66ff0aa50>"
            ]
          },
          "metadata": {},
          "execution_count": 118
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Evaluate turkey model on test data\n",
        "\n",
        "turkey_model.evaluate(X_test, y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "N_ZhAeFZ3PtU",
        "outputId": "651ab7a4-9895-4277-c387-1f144f3bbb0e"
      },
      "execution_count": 119,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "18/18 [==============================] - 0s 3ms/step - loss: 807.1151\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "807.1150512695312"
            ]
          },
          "metadata": {},
          "execution_count": 119
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Load best performing turkey model and evaluate on test data\n",
        "\n",
        "turkey_model = tf.keras.models.load_model(filepath = '/content/drive/MyDrive/BTC Predictor/model_experiments/model_6_turkey_model/')\n",
        "turkey_model.evaluate(X_test, y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eyYvzkgk5goa",
        "outputId": "8810dd73-66f0-480e-ce67-712f6889914f"
      },
      "execution_count": 120,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "18/18 [==============================] - 0s 3ms/step - loss: 637.4183\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "637.4182739257812"
            ]
          },
          "metadata": {},
          "execution_count": 120
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Make predictions with turkey model\n",
        "\n",
        "turkey_preds = make_preds(turkey_model, X_test)\n",
        "turkey_preds[:10]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DFo91B-_53mU",
        "outputId": "588f7b91-15f9-40bf-f674-f68ca6bab5d0"
      },
      "execution_count": 121,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Tensor: shape=(10,), dtype=float32, numpy=\n",
              "array([8817.066 , 8783.974 , 9029.472 , 8763.059 , 8774.349 , 8712.79  ,\n",
              "       8650.94  , 8553.647 , 8434.3125, 8527.643 ], dtype=float32)>"
            ]
          },
          "metadata": {},
          "execution_count": 121
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Evaluate turkey preds\n",
        "\n",
        "turkey_results = evaluate_preds(y_true = y_test, \n",
        "                                y_pred = turkey_preds)\n",
        "turkey_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-SSrlR6F6YgM",
        "outputId": "0c56f3ba-29f3-424b-f463-a95a98f692b3"
      },
      "execution_count": 122,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'mae': 17134.238,\n",
              " 'mape': 121.42651,\n",
              " 'mase': 26.515285,\n",
              " 'mse': 614681100.0,\n",
              " 'rmse': 23721.217}"
            ]
          },
          "metadata": {},
          "execution_count": 122
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model_1_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DjBE5iQ-6g1C",
        "outputId": "d76017b2-cc92-4d92-b7c0-d159e24862be"
      },
      "execution_count": 123,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'mae': 568.95105,\n",
              " 'mape': 2.5448983,\n",
              " 'mase': 0.9994894,\n",
              " 'mse': 1171743.8,\n",
              " 'rmse': 1082.4711}"
            ]
          },
          "metadata": {},
          "execution_count": 123
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Plotting the results\n",
        "\n",
        "plt.figure(figsize = (10, 7))\n",
        "offset = 0\n",
        "plot_time_series(timesteps = btc_timesteps_turkey[-len(X_test):], \n",
        "                 values = btc_price_turkey[-len(y_test):],\n",
        "                 format = '-',\n",
        "                 start = offset,\n",
        "                 label = \"Turkey Test Data\")\n",
        "\n",
        "plot_time_series(timesteps = btc_timesteps_turkey[-len(X_test):], \n",
        "                 values = turkey_preds,\n",
        "                 format = '-',\n",
        "                 start = offset,\n",
        "                 label = \"Turkey Preds\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 442
        },
        "id": "AG7Ixc8x66_K",
        "outputId": "45c112c1-c80b-4035-c601-e11044874f17"
      },
      "execution_count": 124,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x504 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Why does our model fail to capture the turkey problem data point?\n",
        "\n",
        "Think about it like this, just like a turkey who lives 1000 joyful days, based on observation alone has no reason to believe day 1001 won't be as joyful as the last, a model which has been trained on historical data of Bitcoin which has no single event where the price decreased by 100x in a day, has no reason to predict it will in the future.\n",
        "\n",
        "A model cannot predict anything in the future outside of the distribution it was trained on.\n",
        "\n",
        "In turn, highly unlikely price movements (based on historical movements), upward or downward will likely never be part of a forecast.\n",
        "\n",
        "However, as we've seen, despite their unlikeliness, these events can have catastrophic impacts to the performance of our models."
      ],
      "metadata": {
        "id": "osQXpBqM9LYl"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Comparing Models"
      ],
      "metadata": {
        "id": "vqZE4rbM9max"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Compare different model resutls (w = window, h = horizon, e.g. w = 7 means a window_size of 7)\n",
        "\n",
        "model_results = pd.DataFrame({\"naive_model\": naive_results,\n",
        "                             \"model_1_dense_w7_h1\": model_1_results,\n",
        "                             \"model_2_dense_w30_h1\": model_2_results,\n",
        "                             \"model_3_dense_w30_h7\": model_3_results,\n",
        "                             \"model_4_LSTM\": model_4_results,\n",
        "                             \"model_5_ensemble\": ensemble_results,\n",
        "                             \"model_6_turkey_model\": turkey_results}).T\n",
        "              \n",
        "model_results"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        },
        "id": "_ve-64Jl7ebF",
        "outputId": "ac7b1c24-f9c5-4205-b86c-7f8295eb9092"
      },
      "execution_count": 125,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                               mae           mse          rmse        mape  \\\n",
              "naive_model             567.980225  1.147547e+06   1071.236206    2.516525   \n",
              "model_1_dense_w7_h1     568.951050  1.171744e+06   1082.471069    2.544898   \n",
              "model_2_dense_w30_h1    605.874084  1.273361e+06   1128.432861    2.726445   \n",
              "model_3_dense_w30_h7   1237.506470  5.405199e+06   1425.747803    5.558879   \n",
              "model_4_LSTM            594.602966  1.267232e+06   1125.713867    2.671757   \n",
              "model_5_ensemble        565.047913  1.131962e+06   1063.936768    2.554046   \n",
              "model_6_turkey_model  17134.238281  6.146811e+08  23721.216797  121.426514   \n",
              "\n",
              "                           mase  \n",
              "naive_model            0.999570  \n",
              "model_1_dense_w7_h1    0.999489  \n",
              "model_2_dense_w30_h1   1.059074  \n",
              "model_3_dense_w30_h7   2.202073  \n",
              "model_4_LSTM           1.044553  \n",
              "model_5_ensemble       0.992633  \n",
              "model_6_turkey_model  26.515285  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-94258da3-2a2e-4d27-ba2d-8ede8f6dff6d\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>mae</th>\n",
              "      <th>mse</th>\n",
              "      <th>rmse</th>\n",
              "      <th>mape</th>\n",
              "      <th>mase</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>naive_model</th>\n",
              "      <td>567.980225</td>\n",
              "      <td>1.147547e+06</td>\n",
              "      <td>1071.236206</td>\n",
              "      <td>2.516525</td>\n",
              "      <td>0.999570</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>model_1_dense_w7_h1</th>\n",
              "      <td>568.951050</td>\n",
              "      <td>1.171744e+06</td>\n",
              "      <td>1082.471069</td>\n",
              "      <td>2.544898</td>\n",
              "      <td>0.999489</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>model_2_dense_w30_h1</th>\n",
              "      <td>605.874084</td>\n",
              "      <td>1.273361e+06</td>\n",
              "      <td>1128.432861</td>\n",
              "      <td>2.726445</td>\n",
              "      <td>1.059074</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>model_3_dense_w30_h7</th>\n",
              "      <td>1237.506470</td>\n",
              "      <td>5.405199e+06</td>\n",
              "      <td>1425.747803</td>\n",
              "      <td>5.558879</td>\n",
              "      <td>2.202073</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>model_4_LSTM</th>\n",
              "      <td>594.602966</td>\n",
              "      <td>1.267232e+06</td>\n",
              "      <td>1125.713867</td>\n",
              "      <td>2.671757</td>\n",
              "      <td>1.044553</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>model_5_ensemble</th>\n",
              "      <td>565.047913</td>\n",
              "      <td>1.131962e+06</td>\n",
              "      <td>1063.936768</td>\n",
              "      <td>2.554046</td>\n",
              "      <td>0.992633</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>model_6_turkey_model</th>\n",
              "      <td>17134.238281</td>\n",
              "      <td>6.146811e+08</td>\n",
              "      <td>23721.216797</td>\n",
              "      <td>121.426514</td>\n",
              "      <td>26.515285</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-94258da3-2a2e-4d27-ba2d-8ede8f6dff6d')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-94258da3-2a2e-4d27-ba2d-8ede8f6dff6d button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-94258da3-2a2e-4d27-ba2d-8ede8f6dff6d');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 125
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Sort model results by MAE and plot them\n",
        "\n",
        "model_results = model_results[[\"mae\"]].sort_values(by=\"mae\")\n",
        "\n",
        "model_results.plot(figsize=(10, 7), kind = \"bar\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 557
        },
        "id": "khwxlQas_e4E",
        "outputId": "2f4347cb-3c3b-44cd-eec3-bab1e2738e80"
      },
      "execution_count": 135,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7fc65e96ea50>"
            ]
          },
          "metadata": {},
          "execution_count": 135
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x504 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model_results.plot.pie(y=\"mae\", figsize = (10,7), colors = ['red', 'pink', 'blue', 'green', 'magenta', 'orange', 'cyan'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 429
        },
        "id": "PpxdDRqqAEAf",
        "outputId": "ba45dfcb-50fb-460a-a54d-047ace29ad6d"
      },
      "execution_count": 143,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7fc65ceaaf50>"
            ]
          },
          "metadata": {},
          "execution_count": 143
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x504 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "\n",
        "* Only the ensemble model performs on par or only slightly better than the naive model. And for the turkey model, changing a single data point destroys its performance.\n",
        "\n",
        "* Forecasting in a closed system (such as predicting the demand of electricity) often yields quite usable results, while forecasting methods usually don't perform well in open systems.\n",
        "\n",
        "* So not knowing the future is a benefit. A known future is already the past.\n"
      ],
      "metadata": {
        "id": "ttFdmzgeB6i7"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# THE END"
      ],
      "metadata": {
        "id": "ub7LWSrWMMlh"
      }
    }
  ]
}