homework/Day9_homework_embeddings_v2.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "***\n",
    "***\n",
    "\n",
    "## The Geometry of Culture\n",
    "\n",
    "Analyzing Meaning through Word Embeddings\n",
    "\n",
    "***\n",
    "***\n",
    "\n",
    "Code and data associated with the ASR paper on the Geometry of Culture. The full paper can be found here: https://journals.sagepub.com/doi/full/10.1177/0003122419877135\n",
    "\n",
    "Austin C. Kozlowski; Matt Taddy; James A. Evans\n",
    "\n",
    "Word embeddings represent **semantic relations** between words as **geometric relationships** between vectors in a high-dimensional space, operationalizing a relational model of meaning consistent with contemporary theories of identity and culture. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "## Download this repo\n",
    "https://github.com/KnowledgeLab/GeometryofCulture"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "\n",
    "- Dimensions induced by word differences (e.g. man - woman, rich - poor, black - white, liberal - conservative) in these vector spaces closely correspond to dimensions of cultural meaning, \n",
    "- Macro-cultural investigation with a longitudinal analysis of the coevolution of gender and class associations in the United States over the 20th century \n",
    "\n",
    "The success of these high-dimensional models motivates a move towards \"high-dimensional theorizing\" of meanings, identities and cultural processes."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "$$\n",
    "\\text{similarity} = \\cos(\\theta) = {\\mathbf{A} \\cdot \\mathbf{B} \\over \\|\\mathbf{A}\\| \\|\\mathbf{B}\\|} = \\frac{ \\sum\\limits_{i=1}^{n}{A_i  B_i} }{ \\sqrt{\\sum\\limits_{i=1}^{n}{A_i^2}}  \\sqrt{\\sum\\limits_{i=1}^{n}{B_i^2}} },\n",
    "$$\n",
    "\n",
    "where $A_i$ and $B_i$ are components of vector $A$ and $B$ respectively."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T05:04:59.325063Z",
     "start_time": "2020-08-14T05:04:59.318970Z"
    },
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-0.4999999999999999"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from numpy import dot\n",
    "from numpy.linalg import norm\n",
    "\n",
    "def cos_sim(a, b):\n",
    "    return dot(a, b)/(norm(a)*norm(b))\n",
    "\n",
    "cos_sim([1, 0, -1], [-1,-1, 0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T05:05:01.584412Z",
     "start_time": "2020-08-14T05:05:01.439890Z"
    },
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-0.5]])"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics.pairwise import cosine_similarity\n",
    "\n",
    "cosine_similarity([[1, 0, -1]], [[-1,-1, 0]])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "### Google News Word2Vec\n",
    "\n",
    "You can download Google’s pre-trained model here.\n",
    "\n",
    "- It’s 1.5GB! \n",
    "- It includes word vectors for a vocabulary of 3 million words and phrases \n",
    "- It is trained on roughly 100 billion words from a Google News dataset. \n",
    "- The vector length is 300 features.\n",
    "\n",
    "### Download GoogleNews-vectors-negative300.bin\n",
    "\n",
    "- intro: http://mccormickml.com/2016/04/12/googles-pretrained-word2vec-model-in-python/\n",
    "- intro: https://code.google.com/archive/p/word2vec/\n",
    "- download: https://drive.google.com/file/d/0B7XkCwpI5KDYNlNUTTlSS21pQmM/edit?usp=sharing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "Using the **Gensim** library in python, we can \n",
    "- find the most similar words to the resulting vector. \n",
    "- add and subtract word vectors, \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T03:46:06.885174Z",
     "start_time": "2020-08-14T03:45:43.187009Z"
    },
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [],
   "source": [
    "import gensim\n",
    "# Load Google's pre-trained Word2Vec model.\n",
    "filepath = '/Users/datalab/bigdata/GoogleNews-vectors-negative300.bin'\n",
    "model = gensim.models.KeyedVectors.load_word2vec_format(filepath, binary=True) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T03:46:24.610657Z",
     "start_time": "2020-08-14T03:46:24.606517Z"
    },
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.24316406, -0.07714844, -0.10302734, -0.10742188,  0.11816406,\n",
       "       -0.10742188, -0.11425781,  0.02563477,  0.11181641,  0.04858398],\n",
       "      dtype=float32)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model['woman'][:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T03:48:06.918378Z",
     "start_time": "2020-08-14T03:48:06.713168Z"
    },
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('Plato_Apology', 0.5661361217498779),\n",
       " ('philosopher_Socrates', 0.561074435710907),\n",
       " ('Aristotle', 0.5599992275238037),\n",
       " ('Parmenides', 0.541626513004303),\n",
       " ('Sophists', 0.5381807088851929),\n",
       " ('Sócrates', 0.533310055732727),\n",
       " ('Plato', 0.5317621827125549),\n",
       " ('philosopher_Aristotle', 0.5256754159927368),\n",
       " ('Plato_Socrates', 0.524064838886261),\n",
       " ('philosopher_Plato', 0.5164784789085388)]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.most_similar('Socrates')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T03:49:03.222758Z",
     "start_time": "2020-08-14T03:49:03.218521Z"
    },
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.76640123"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.similarity('woman', 'man')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T05:05:13.891241Z",
     "start_time": "2020-08-14T05:05:13.887119Z"
    },
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.76640123"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cos_sim(model['woman'], model['man'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T05:05:18.768602Z",
     "start_time": "2020-08-14T05:05:17.725021Z"
    },
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('queen', 0.7118192911148071),\n",
       " ('monarch', 0.6189674139022827),\n",
       " ('princess', 0.5902431607246399),\n",
       " ('crown_prince', 0.5499460697174072),\n",
       " ('prince', 0.5377321243286133)]"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.most_similar(positive=['woman', 'king'], negative=['man'], topn=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T05:35:16.629282Z",
     "start_time": "2020-08-14T05:35:16.624343Z"
    },
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [],
   "source": [
    "from numpy import dot  #Dot product#\n",
    "from numpy.linalg import norm\n",
    "from sklearn.metrics.pairwise import cosine_similarity\n",
    "from scipy import spatial\n",
    "\n",
    "\n",
    "def cos_sim(a, b):\n",
    "    return dot(a, b)/(norm(a)*norm(b))\n",
    "\n",
    "def cos(x, y):\n",
    "    return 1 - spatial.distance.cosine(x, y)\n",
    "\n",
    "def nrm(x):\n",
    "    # Normalize vector#\n",
    "    return x/norm(x)\n",
    "\n",
    "def dimension(x, y):\n",
    "    # Calculate semantic dimension from antonym pair#\n",
    "    return nrm(nrm(x)-nrm(y))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T05:35:18.683800Z",
     "start_time": "2020-08-14T05:35:18.674726Z"
    },
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [],
   "source": [
    "###IMPORT LISTS OF TERMS TO PROJECT AND ANTONYM PAIRS#####\n",
    "import pandas as pd\n",
    "ant_pairs_aff = pd.read_csv(\"../data/word_pairs/affluence_pairs.csv\",header=None)\n",
    "ant_pairs_gen = pd.read_csv()  # change here\n",
    "ant_pairs_race = pd.read_csv() # change here"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T05:43:51.808816Z",
     "start_time": "2020-08-14T05:43:51.804442Z"
    },
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [],
   "source": [
    "def make_dim(model, df):\n",
    "    word_dims = []\n",
    "    for i in df.index:\n",
    "        rp_word1 = df[0][i]\n",
    "        rp_word2 = df[1][i]\n",
    "        try:\n",
    "            # 计算rp_word1、rp_word2两个词向量构成的dimension\n",
    "            word_dims.append( ) # change here\n",
    "        except:\n",
    "            pass\n",
    "    dim_ave = pd.DataFrame(word_dims).mean().tolist()\n",
    "    dim_ave_= nrm(dim_ave)\n",
    "    return dim_ave\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T05:45:48.352084Z",
     "start_time": "2020-08-14T05:45:48.298841Z"
    },
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [],
   "source": [
    "aff_dim = make_dim(model, ant_pairs_aff)\n",
    "gender_dim = make_dim( ) # change here\n",
    "race_dim = make_dim( ) # change here\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T05:45:59.641849Z",
     "start_time": "2020-08-14T05:45:59.637310Z"
    },
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(-0.014282578052009498, -0.10834102639266807, -0.06033891833467253)"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 计算aff_dim、gender_dim、race_dim三者的cosine similarity\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T05:53:15.147398Z",
     "start_time": "2020-08-14T05:53:15.141404Z"
    },
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [],
   "source": [
    "wlist=[\"camping\",\"baseball\",\"boxing\",\"volleyball\",\"softball\",\"golf\",\"tennis\",\"soccer\",\"basketball\",\"hockey\"]\n",
    "\n",
    "aff_proj = [cos(aff_dim, model[i]) for i in wlist]\n",
    "gender_proj = [] # change here\n",
    "race_proj = [ ]# change here"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T05:53:33.837101Z",
     "start_time": "2020-08-14T05:53:33.832589Z"
    },
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.07256063,  0.12535858,  0.07870991,  0.0066219 , -0.00821159],\n",
       "      dtype=float32)"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dimension(model['man'], model['woman'])[:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T05:58:55.051974Z",
     "start_time": "2020-08-14T05:58:54.882736Z"
    },
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import matplotlib\n",
    "matplotlib.style.use('ggplot')\n",
    "\n",
    "# 绘图"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T06:00:48.733815Z",
     "start_time": "2020-08-14T06:00:48.574819Z"
    },
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 绘图\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "Now that we’ve looked at trained word embeddings, \n",
    "\n",
    "- let’s learn more about the training process. \n",
    "- But before we get to word2vec, we need to look at a conceptual parent of word embeddings: **the neural language model**.\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Gensim Word2vec "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T06:38:40.740911Z",
     "start_time": "2020-08-14T06:38:39.309264Z"
    },
    "slideshow": {
     "slide_type": "fragment"
    }
   },
   "outputs": [],
   "source": [
    "import gensim as gensim\n",
    "from gensim.models import Word2Vec\n",
    "from gensim.models.keyedvectors import KeyedVectors\n",
    "from gensim.models.word2vec import LineSentence"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T06:49:16.828954Z",
     "start_time": "2020-08-14T06:49:16.821352Z"
    },
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [],
   "source": [
    "# 下载理想国英文版本txt文件，网址为：http://www.gutenberg.org/cache/epub/150/pg150.txt\n",
    "\n",
    "with open(\"pg150.txt\", 'r') as f:\n",
    "    lines = f.readlines()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T06:49:17.390106Z",
     "start_time": "2020-08-14T06:49:17.386315Z"
    },
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "16960"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(lines)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T06:49:33.192579Z",
     "start_time": "2020-08-14T06:49:33.180342Z"
    },
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [],
   "source": [
    "book = []\n",
    "paragraph = ''\n",
    "for i in lines:\n",
    "    if i !='\\n':\n",
    "        paragraph = paragraph + ' '+ i\n",
    "    else:\n",
    "        book.append(paragraph)\n",
    "        paragraph = ''\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T06:49:33.668616Z",
     "start_time": "2020-08-14T06:49:33.664271Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4519"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(book)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T06:49:34.238414Z",
     "start_time": "2020-08-14T06:49:34.219461Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['',\n",
       " '\\ufeffThe',\n",
       " 'Project',\n",
       " 'Gutenberg',\n",
       " 'EBook',\n",
       " 'of',\n",
       " \"Plato's\",\n",
       " 'Republic,',\n",
       " 'by',\n",
       " 'Plato']"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "book = [i.replace('\\n', '').split(' ') for i in book]\n",
    "\n",
    "book[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T06:49:43.875182Z",
     "start_time": "2020-08-14T06:49:40.739848Z"
    },
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [],
   "source": [
    "# 调用gensim Word2Vec的算法进行训练。\n",
    "# 参数分别为：size: 100 嵌入后的词向量维度；window: 2上下文的宽度，min_count 0 为考虑计算的单词的最低词频阈值\n",
    "model = Word2Vec( ) # change here"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T06:49:55.113272Z",
     "start_time": "2020-08-14T06:49:55.107747Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:1: DeprecationWarning: Call to deprecated `__getitem__` (Method will be removed in 4.0.0, use self.wv.__getitem__() instead).\n",
      "  \"\"\"Entry point for launching an IPython kernel.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([-1.37807399e-01, -8.50972980e-02, -1.13084763e-02, -5.47388643e-02,\n",
       "       -7.65230134e-03,  4.63625900e-02, -1.47486730e-02,  5.45147732e-02,\n",
       "       -4.95448150e-02,  3.51702049e-02, -2.01568142e-01, -2.71834917e-02,\n",
       "        2.41141934e-02, -2.37497687e-03,  5.85087836e-02, -2.91526467e-02,\n",
       "       -3.89404669e-02, -3.37465070e-02,  3.65749188e-02, -3.33004259e-02,\n",
       "        3.11295446e-02, -4.64655086e-02, -2.74411552e-02, -1.17913283e-01,\n",
       "       -9.69972741e-03, -2.03968666e-04,  3.42998542e-02, -4.59230319e-02,\n",
       "       -2.28419363e-01,  7.53095001e-02,  1.16067998e-01,  2.36383397e-02,\n",
       "        1.27256542e-01, -7.05621243e-02,  1.00037888e-01,  4.58286814e-02,\n",
       "       -5.83368801e-02, -8.03394169e-02, -1.82675138e-01, -6.63211495e-02,\n",
       "        8.43277853e-03, -6.84499145e-02,  1.28329352e-01, -1.21511919e-02,\n",
       "        3.29800136e-02,  4.06918265e-02,  6.52354583e-02,  1.52770486e-02,\n",
       "       -1.01396181e-01,  1.15474589e-01, -7.89184794e-02,  3.35669294e-02,\n",
       "        6.52226880e-02, -6.70596510e-02, -4.89763245e-02,  1.98891237e-01,\n",
       "        6.48066774e-02, -5.29073626e-02,  1.17686436e-01,  6.05571494e-02,\n",
       "        4.81364317e-03,  1.12978658e-02, -1.21227279e-01, -8.13504159e-02,\n",
       "        1.93354748e-02,  8.40817466e-02, -5.93261458e-02,  1.06367692e-01,\n",
       "       -1.17695704e-01,  1.33523375e-01, -1.05052598e-01,  8.11126828e-02,\n",
       "       -4.79990542e-02, -4.17293869e-02,  7.54278749e-02, -6.26239702e-02,\n",
       "        2.10881862e-03, -6.99369609e-02, -3.48991528e-02, -1.98208854e-01,\n",
       "        8.24960545e-02, -3.20518985e-02, -1.70636520e-01,  3.31224762e-02,\n",
       "       -1.96487531e-01,  3.06039937e-02,  5.15202433e-03, -1.20531274e-02,\n",
       "       -3.57166678e-02,  5.36215603e-02,  5.61406575e-02,  1.57578033e-03,\n",
       "        2.71171853e-02, -1.40045956e-01, -7.39080384e-02,  2.55121794e-02,\n",
       "       -9.88911688e-02, -7.16031566e-02, -8.72703791e-02,  1.23933800e-01],\n",
       "      dtype=float32)"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model['Socrates']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T06:50:04.021580Z",
     "start_time": "2020-08-14T06:50:04.003200Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('words', 0.9993601441383362),\n",
       " ('her', 0.9993101954460144),\n",
       " ('always', 0.9993042945861816),\n",
       " ('tyrant', 0.9992969036102295),\n",
       " ('real', 0.999282717704773),\n",
       " ('injustice', 0.9992785453796387),\n",
       " ('himself', 0.9992719888687134),\n",
       " ('public', 0.9992649555206299),\n",
       " ('after', 0.999255895614624),\n",
       " ('reason', 0.9992557764053345)]"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.wv.most_similar('Socrates', topn = 10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T06:54:37.439056Z",
     "start_time": "2020-08-14T06:54:37.433219Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('best', 0.9997957944869995),\n",
       " ('some', 0.9997676610946655),\n",
       " ('use', 0.9997552633285522),\n",
       " ('opinion', 0.9997509121894836),\n",
       " ('another', 0.9997390508651733),\n",
       " ('either', 0.9997327327728271),\n",
       " ('our', 0.9997314214706421),\n",
       " ('evil', 0.9997271299362183),\n",
       " ('pleasure', 0.999714195728302),\n",
       " ('thing', 0.9997127652168274)]"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.wv.most_similar('justice', topn = 10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T06:56:59.439904Z",
     "start_time": "2020-08-14T06:56:59.434075Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('good', 0.9996432065963745),\n",
       " ('city', 0.9995303153991699),\n",
       " ('him', 0.999497652053833),\n",
       " ('another', 0.9994549751281738),\n",
       " ('life', 0.9994546175003052),\n",
       " ('State,', 0.9994027018547058),\n",
       " ('way', 0.9993818998336792),\n",
       " ('evil', 0.9993690848350525),\n",
       " ('time', 0.999363899230957),\n",
       " ('class', 0.9993453025817871)]"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.wv.most_similar('man', topn = 10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T06:50:52.526514Z",
     "start_time": "2020-08-14T06:50:52.123980Z"
    },
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:8: DeprecationWarning: Call to deprecated `__getitem__` (Method will be removed in 4.0.0, use self.wv.__getitem__() instead).\n",
      "  \n"
     ]
    }
   ],
   "source": [
    "# 将词向量投影到二维空间\n",
    "import numpy as np\n",
    "from sklearn.decomposition import PCA\n",
    "\n",
    "rawWordVec = []\n",
    "word2ind = {}\n",
    "for i, w in enumerate(model.wv.vocab):\n",
    "    rawWordVec.append(model[w])\n",
    "    word2ind[w] = i\n",
    "rawWordVec = np.array(rawWordVec)\n",
    "X_reduced = PCA( ).fit_transform(rawWordVec) # change here"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-08-14T06:58:51.837481Z",
     "start_time": "2020-08-14T06:58:51.578618Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 绘制星空图\n",
    "# 绘制所有单词向量的二维空间投影\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "fig = plt.figure(figsize = (15, 10))\n",
    "ax = fig.gca()\n",
    "ax.set_facecolor('black')\n",
    "ax.plot(X_reduced[:, 0], X_reduced[:, 1], '.', markersize = 1, alpha = 0.5, color = 'white')\n",
    "# 绘制几个特殊单词的向量\n",
    "words = [] # change here\n",
    "for w in words:\n",
    "    if w in word2ind:\n",
    "        ind = word2ind[w]\n",
    "        xy = X_reduced[ind]\n",
    "        plt.plot(xy[0], xy[1], '.', alpha =1, color = 'red')\n",
    "        plt.text(xy[0], xy[1], w,  alpha = 1, color = 'yellow')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# END"
   ]
  }
 ],
 "metadata": {
  "celltoolbar": "Slideshow",
  "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": "3.9.7"
  },
  "latex_envs": {
   "LaTeX_envs_menu_present": true,
   "autoclose": false,
   "autocomplete": true,
   "bibliofile": "biblio.bib",
   "cite_by": "apalike",
   "current_citInitial": 1,
   "eqLabelWithNumbers": true,
   "eqNumInitial": 1,
   "hotkeys": {
    "equation": "Ctrl-E",
    "itemize": "Ctrl-I"
   },
   "labels_anchors": false,
   "latex_user_defs": false,
   "report_style_numbering": false,
   "user_envs_cfg": false
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": false,
   "sideBar": false,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {
    "height": "517px",
    "left": "953px",
    "top": "134px",
    "width": "255px"
   },
   "toc_section_display": false,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}