new-house-CNN-attempt1.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "3d299c66",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import os\n",
    "from PIL import Image\n",
    "\n",
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "from tensorflow.keras import Sequential\n",
    "from tensorflow.keras.layers import Dense, Conv2D, Dropout, Flatten, MaxPooling2D\n",
    "from tensorflow.keras.models import model_from_json\n",
    "\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.metrics import accuracy_score as accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "653a118f",
   "metadata": {},
   "outputs": [],
   "source": [
    "def toy_CNN(X_train, y_train, **kwargs):\n",
    "        \n",
    "    model = Sequential([\n",
    "        Conv2D(filters=8, kernel_size=(8,8), strides=(4,4), activation='relu', input_shape=(384, 512, 3)),\n",
    "        MaxPooling2D(pool_size=(8,8)),\n",
    "        Conv2D(filters=8, kernel_size=(8,8), strides=(4,4), activation='relu'),\n",
    "        Dropout(0.1),\n",
    "        Flatten(),\n",
    "        Dense(1, activation='sigmoid')\n",
    "    ])\n",
    "    \n",
    "    model.compile(loss='binary_crossentropy', optimizer='adam', metrics='accuracy')\n",
    "    history = model.fit(X_train, y_train, **kwargs)\n",
    "    return model, history"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "3c834abd",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Available datapoints: 352\n",
      "................................................................................................................................................................................................................................................................................................................................................................\n",
      "Kept datapoints: 352\n",
      "Trash datapoints: 118\n",
      "done\n"
     ]
    }
   ],
   "source": [
    "X, y = [], []\n",
    "folder = 'new-house-resized-dataset'\n",
    "folder_list = os.listdir(folder)\n",
    "nTotal = len(folder_list)\n",
    "print('Available datapoints:', nTotal)\n",
    "nTrash = 0\n",
    "for file in folder_list:\n",
    "    print('.',end='')\n",
    "    trash = int(file.split('_')[0] == 'trash')\n",
    "    img = Image.open(f'{folder}/{file}')\n",
    "    arr = np.array(img)\n",
    "    X.append(arr)\n",
    "    y.append(trash)\n",
    "    nTrash += trash\n",
    "\n",
    "print()\n",
    "print('Kept datapoints:', len(y))\n",
    "print('Trash datapoints:', nTrash)\n",
    "X, y = np.array(X), np.array(y)\n",
    "print('done')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "d8a6f920",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " -- Performing NEW train-test split -- \n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(70, 384, 512, 3)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def train_test_split(X, y, test_idx=None, test_size=0.2):\n",
    "    ''' Split input data X and labels y into training and testing arrays '''\n",
    "    n_list = list(range(len(y)))\n",
    "    if test_idx is None:\n",
    "        print('\\n -- Performing NEW train-test split -- \\n')\n",
    "        test_idx = np.random.choice(n_list, replace=False, size=int(test_size*len(y)))\n",
    "    else:\n",
    "        test_idx = np.array(test_idx)\n",
    "    train_idx = np.array([i for i in n_list if i not in test_idx])\n",
    "    X_train = X[train_idx,:]\n",
    "    X_test = X[test_idx,:]\n",
    "    y_train = y[train_idx]\n",
    "    y_test = y[test_idx]\n",
    "    return X_train, y_train, X_test, y_test, test_idx\n",
    "\n",
    "X_train, y_train, X_test, y_test, test_idx = train_test_split(X,y)\n",
    "X_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "381aadb8",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/20\n",
      "8/8 [==============================] - 2s 201ms/step - loss: 18.0014 - accuracy: 0.7244 - val_loss: 3.8795 - val_accuracy: 0.7544\n",
      "Epoch 2/20\n",
      "8/8 [==============================] - 1s 137ms/step - loss: 2.6821 - accuracy: 0.7244 - val_loss: 11.7224 - val_accuracy: 0.1053\n",
      "Epoch 3/20\n",
      "8/8 [==============================] - 1s 136ms/step - loss: 1.1252 - accuracy: 0.8533 - val_loss: 3.9710 - val_accuracy: 0.2105\n",
      "Epoch 4/20\n",
      "8/8 [==============================] - 1s 133ms/step - loss: 0.3505 - accuracy: 0.8711 - val_loss: 1.6350 - val_accuracy: 0.3509\n",
      "Epoch 5/20\n",
      "8/8 [==============================] - 1s 143ms/step - loss: 0.2281 - accuracy: 0.9200 - val_loss: 0.9758 - val_accuracy: 0.4912\n",
      "Epoch 6/20\n",
      "8/8 [==============================] - 1s 137ms/step - loss: 0.2083 - accuracy: 0.9244 - val_loss: 1.1625 - val_accuracy: 0.4211\n",
      "Epoch 7/20\n",
      "8/8 [==============================] - 1s 144ms/step - loss: 0.1635 - accuracy: 0.9378 - val_loss: 1.4032 - val_accuracy: 0.3684\n",
      "Epoch 8/20\n",
      "8/8 [==============================] - 1s 141ms/step - loss: 0.1376 - accuracy: 0.9644 - val_loss: 1.3182 - val_accuracy: 0.4386\n",
      "Epoch 9/20\n",
      "8/8 [==============================] - 1s 142ms/step - loss: 0.1238 - accuracy: 0.9556 - val_loss: 1.2271 - val_accuracy: 0.4386\n",
      "Epoch 10/20\n",
      "8/8 [==============================] - 1s 141ms/step - loss: 0.1042 - accuracy: 0.9733 - val_loss: 0.9987 - val_accuracy: 0.5789\n",
      "Epoch 11/20\n",
      "8/8 [==============================] - 1s 143ms/step - loss: 0.1060 - accuracy: 0.9822 - val_loss: 1.0407 - val_accuracy: 0.7368\n",
      "Epoch 12/20\n",
      "8/8 [==============================] - 1s 141ms/step - loss: 0.1123 - accuracy: 0.9600 - val_loss: 1.0430 - val_accuracy: 0.6491\n",
      "Epoch 13/20\n",
      "8/8 [==============================] - 1s 146ms/step - loss: 0.0661 - accuracy: 0.9778 - val_loss: 0.9740 - val_accuracy: 0.6842\n",
      "Epoch 14/20\n",
      "8/8 [==============================] - 1s 145ms/step - loss: 0.0871 - accuracy: 0.9644 - val_loss: 1.1645 - val_accuracy: 0.7368\n",
      "Epoch 15/20\n",
      "8/8 [==============================] - 1s 142ms/step - loss: 0.0772 - accuracy: 0.9733 - val_loss: 0.8173 - val_accuracy: 0.7368\n",
      "Epoch 16/20\n",
      "8/8 [==============================] - 1s 142ms/step - loss: 0.0632 - accuracy: 0.9778 - val_loss: 0.9924 - val_accuracy: 0.6667\n",
      "Epoch 17/20\n",
      "8/8 [==============================] - 1s 146ms/step - loss: 0.0818 - accuracy: 0.9733 - val_loss: 0.9777 - val_accuracy: 0.6667\n",
      "Epoch 18/20\n",
      "8/8 [==============================] - 1s 133ms/step - loss: 0.0770 - accuracy: 0.9600 - val_loss: 0.8156 - val_accuracy: 0.7193\n",
      "Epoch 19/20\n",
      "8/8 [==============================] - 1s 140ms/step - loss: 0.0517 - accuracy: 0.9733 - val_loss: 0.6696 - val_accuracy: 0.7368\n",
      "Epoch 20/20\n",
      "8/8 [==============================] - 1s 148ms/step - loss: 0.0658 - accuracy: 0.9822 - val_loss: 0.7088 - val_accuracy: 0.7719\n"
     ]
    }
   ],
   "source": [
    "#class_weight = {0:1, 1:100}\n",
    "#print(class_weight)\n",
    "toy_model, toy_hist = toy_CNN(X_train, y_train, verbose=True, epochs=20, validation_split=0.2,)\n",
    "                                # class_weight = class_weight)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "3cec5c0b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plot_loss(hist):\n",
    "    plt.figure()\n",
    "    plt.plot(hist.history['loss'], label='loss')\n",
    "    plt.plot(hist.history['val_loss'], label='val_loss')\n",
    "    plt.xlabel('Epoch')\n",
    "    plt.ylabel('Loss (Binary Cross Entropy)')\n",
    "    plt.legend()\n",
    "    plt.grid(True)\n",
    "    \n",
    "plot_loss(toy_hist)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "140f33f6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_pred = np.round(toy_model.predict(X_test), 4)\n",
    "plt.figure(figsize=(6,6))\n",
    "plt.plot(y_test, y_pred, 'o', label='Test Data', ms=10, alpha=0.2, c='green')\n",
    "plt.axvline(x=0.5, ls='--', c='k')\n",
    "plt.axvline(x=0, ls='-', c='grey', zorder=0, lw=0.5)\n",
    "plt.axvline(x=1, ls='-', c='grey', zorder=0, lw=0.5)\n",
    "plt.axhline(y=0.5, ls='--', c='k')\n",
    "plt.axhline(y=0, ls='-', c='grey', zorder=0, lw=0.5)\n",
    "plt.axhline(y=1, ls='-', c='grey', zorder=0, lw=0.5)\n",
    "plt.xlim(-0.5,1.5)\n",
    "plt.xticks([0,1],['Not Trash', 'Trash'], size=20)\n",
    "plt.ylim(-0.1,1.1)\n",
    "plt.yticks(np.arange(0,1.01,0.1))\n",
    "plt.legend(bbox_to_anchor=(1,1), loc='upper left', edgecolor='k', fontsize=16)\n",
    "plt.title('Prelim CNN test', size=20)\n",
    "None"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "6f0ffdfa",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Trash (1) vs. Non-Trash (0) Binary Classifier trained on New House data\n",
      "\n",
      "Confusion matrix:\n",
      " [[41  2]\n",
      " [10 17]]\n",
      "\n",
      "Accuracy: 0.8285714285714286\n"
     ]
    }
   ],
   "source": [
    "y_pred_round = [int(np.round(a)) for a in y_pred]\n",
    "print('Trash (1) vs. Non-Trash (0) Binary Classifier trained on New House data')\n",
    "print('\\nConfusion matrix:\\n', confusion_matrix(y_test, y_pred_round))\n",
    "print('\\nAccuracy:', accuracy(y_test, y_pred_round))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "c49167bb",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.save('911_X_train', X_train)\n",
    "np.save('911_y_train', y_train)\n",
    "np.save('911_X_test', X_test)\n",
    "np.save('911_y_test', y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "49024e4b",
   "metadata": {},
   "outputs": [],
   "source": [
    "toy_model_json = toy_model.to_json()\n",
    "with open('911_toy_model.json', 'w') as jf:\n",
    "    jf.write(toy_model_json)\n",
    "toy_model.save_weights('911_toy_model.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "6d65a461",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_2 (Conv2D)            (None, 95, 127, 8)        1544      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 11, 15, 8)         0         \n",
      "_________________________________________________________________\n",
      "conv2d_3 (Conv2D)            (None, 1, 2, 8)           4104      \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 1, 2, 8)           0         \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 16)                0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 1)                 17        \n",
      "=================================================================\n",
      "Total params: 5,665\n",
      "Trainable params: 5,665\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# from tensorflow.keras.models import model_from_json\n",
    "with open('911_toy_model.json') as jf:\n",
    "    json = jf.read()\n",
    "loaded_model = model_from_json(json)\n",
    "loaded_model.load_weights('911_toy_model.h5')\n",
    "loaded_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "f18d7db5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_pred = np.round(toy_model.predict(X_test), 4)\n",
    "plt.figure(figsize=(6,6))\n",
    "plt.plot(y_test, y_pred, 'o', label='Test Data', ms=10, alpha=0.2, c='green')\n",
    "plt.axvline(x=0.5, ls='--', c='k')\n",
    "plt.axvline(x=0, ls='-', c='grey', zorder=0, lw=0.5)\n",
    "plt.axvline(x=1, ls='-', c='grey', zorder=0, lw=0.5)\n",
    "plt.axhline(y=0.5, ls='--', c='k')\n",
    "plt.axhline(y=0, ls='-', c='grey', zorder=0, lw=0.5)\n",
    "plt.axhline(y=1, ls='-', c='grey', zorder=0, lw=0.5)\n",
    "plt.xlim(-0.5,1.5)\n",
    "plt.xticks([0,1],['Not Trash', 'Trash'], size=20)\n",
    "plt.ylim(-0.1,1.1)\n",
    "plt.yticks(np.arange(0,1.01,0.1))\n",
    "plt.legend(bbox_to_anchor=(1,1), loc='upper left', edgecolor='k', fontsize=16)\n",
    "plt.title('Prelim CNN test', size=20)\n",
    "None"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "93ef1070",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "arr = np.load('911_X_test.npy')\n",
    "np.sum(arr != X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5da404bb",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}