Training Word Embeddings - Complete Code.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "17f3eea2",
   "metadata": {},
   "source": [
    "### Loading Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "f4e7712f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['The future king is the prince\\n', 'Daughter is the princess\\n', 'Son is the prince\\n', 'Only a man can be a king\\n', 'Only a woman can be a queen\\n', 'The princess will be a queen\\n', 'The prince is a strong man\\n', 'The princess is a beautiful woman\\n', 'Prince is only a boy now\\n', 'Prince will be king\\n', 'A boy will be a man']\n"
     ]
    }
   ],
   "source": [
    "file = open('Royal_data.txt', 'r')\n",
    "royal_data = file.readlines()\n",
    "print(royal_data)\n",
    "file.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1369aa51",
   "metadata": {},
   "source": [
    "### Removing '\\n' from the end of every sentence and convert the sentence into lowercase   \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "88af9298",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['the future king is the prince', 'daughter is the princess', 'son is the prince', 'only a man can be a king', 'only a woman can be a queen', 'the princess will be a queen', 'the prince is a strong man', 'the princess is a beautiful woman', 'prince is only a boy now', 'prince will be king', 'a boy will be a man']\n"
     ]
    }
   ],
   "source": [
    "for i in range(len(royal_data)):\n",
    "    royal_data[i] = royal_data[i].lower().replace('\\n', '')\n",
    "\n",
    "print(royal_data)                                                      \n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "03c1aec2",
   "metadata": {},
   "source": [
    "### Removing stop words\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "236da810",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[['future', 'king', 'prince'], ['daughter', 'princess'], ['son', 'prince'], ['man', 'king'], ['woman', 'queen'], ['princess', 'queen'], ['prince', 'strong', 'man'], ['princess', 'beautiful', 'woman'], ['prince', 'boy'], ['prince', 'king'], ['boy', 'man']]\n"
     ]
    }
   ],
   "source": [
    "stopwords = ['the', 'is', 'will', 'be', 'a', 'only', 'can', 'their', 'now', 'and', 'at', 'it']\n",
    "\n",
    "filtered_data = []\n",
    "for sent in royal_data:\n",
    "    temp = []\n",
    "    for word in sent.split():\n",
    "        if word not in stopwords:\n",
    "            temp.append(word)\n",
    "    filtered_data.append(temp)\n",
    "\n",
    "print(filtered_data)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9c060ff9",
   "metadata": {},
   "source": [
    "### Creating bigrams\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "feb52dc7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[['future', 'king'], ['king', 'future'], ['future', 'prince'], ['prince', 'future'], ['king', 'prince'], ['prince', 'king'], ['daughter', 'princess'], ['princess', 'daughter'], ['son', 'prince'], ['prince', 'son'], ['man', 'king'], ['king', 'man'], ['woman', 'queen'], ['queen', 'woman'], ['princess', 'queen'], ['queen', 'princess'], ['prince', 'strong'], ['strong', 'prince'], ['prince', 'man'], ['man', 'prince'], ['strong', 'man'], ['man', 'strong'], ['princess', 'beautiful'], ['beautiful', 'princess'], ['princess', 'woman'], ['woman', 'princess'], ['beautiful', 'woman'], ['woman', 'beautiful'], ['prince', 'boy'], ['boy', 'prince'], ['prince', 'king'], ['king', 'prince'], ['boy', 'man'], ['man', 'boy']]\n"
     ]
    }
   ],
   "source": [
    "bigrams = []\n",
    "for words_list in filtered_data:\n",
    "    for i in range(len(words_list) - 1):\n",
    "        for j in range(i+1, len(words_list)):\n",
    "            bigrams.append([words_list[i], words_list[j]])\n",
    "            bigrams.append([words_list[j], words_list[i]])\n",
    "            \n",
    "\n",
    "print(bigrams)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a9c73a57",
   "metadata": {},
   "source": [
    "### Getting a list of unique words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "20547a7c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['beautiful', 'boy', 'daughter', 'future', 'king', 'man', 'prince', 'princess', 'queen', 'son', 'strong', 'woman']\n",
      "Total number of unique words are: 12\n"
     ]
    }
   ],
   "source": [
    "all_words = []\n",
    "for bi in bigrams:\n",
    "    all_words.extend(bi)\n",
    "\n",
    "all_words = list(set(all_words))\n",
    "all_words.sort()\n",
    "\n",
    "print(all_words)\n",
    "print(\"Total number of unique words are:\", len(all_words))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2e20adfa",
   "metadata": {},
   "source": [
    "### Creating dictionary of words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "5e8da645",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'beautiful': 0, 'boy': 1, 'daughter': 2, 'future': 3, 'king': 4, 'man': 5, 'prince': 6, 'princess': 7, 'queen': 8, 'son': 9, 'strong': 10, 'woman': 11}\n"
     ]
    }
   ],
   "source": [
    "words_dict = {}\n",
    "\n",
    "counter = 0\n",
    "for word in all_words:\n",
    "    words_dict[word] = counter\n",
    "    counter += 1\n",
    "\n",
    "print(words_dict)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3f911c7d",
   "metadata": {},
   "source": [
    "### Performing one-hot encoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "7062308d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "beautiful : [1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      "boy : [0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      "daughter : [0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      "future : [0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      "king : [0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0.]\n",
      "man : [0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0.]\n",
      "prince : [0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0.]\n",
      "princess : [0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0.]\n",
      "queen : [0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0.]\n",
      "son : [0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0.]\n",
      "strong : [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0.]\n",
      "woman : [0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1.]\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "\n",
    "onehot_data = np.zeros((len(all_words), len(all_words)))\n",
    "\n",
    "for i in range(len(all_words)):\n",
    "    onehot_data[i][i] = 1\n",
    "\n",
    "onehot_dict = {}\n",
    "counter = 0\n",
    "for word in all_words:\n",
    "    onehot_dict[word] = onehot_data[counter]\n",
    "    counter += 1\n",
    "\n",
    "for word in onehot_dict:\n",
    "    print(word, \":\", onehot_dict[word])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "0a6fe9e5",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = []\n",
    "Y = []\n",
    "\n",
    "for bi in bigrams:\n",
    "    X.append(onehot_dict[bi[0]])\n",
    "    Y.append(onehot_dict[bi[1]])\n",
    "    \n",
    "X = np.array(X)\n",
    "Y = np.array(Y)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d218441d",
   "metadata": {},
   "source": [
    "### Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "91f1fa31",
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "\n",
    "embed_size = 2\n",
    "\n",
    "model = Sequential([\n",
    "    Dense(embed_size, activation='linear'),\n",
    "    Dense(Y.shape[1], activation = 'softmax')\n",
    "])\n",
    "\n",
    "model.compile(loss = 'categorical_crossentropy', optimizer = 'adam')\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "3796ca1b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x1811f047700>"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(X, Y, epochs = 1000, batch_size = 256, verbose = False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "2e9d8ebe",
   "metadata": {},
   "outputs": [],
   "source": [
    "weights = model.get_weights()[0]\n",
    "\n",
    "word_embeddings = {}\n",
    "for word in all_words:\n",
    "    word_embeddings[word] = weights[words_dict[word]]\n",
    "\n",
    "# print(word_embeddings)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "95cf5e2d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# plt.figure(figsize = (10, 10))\n",
    "for word in list(words_dict.keys()):\n",
    "    coord = word_embeddings.get(word)\n",
    "    plt.scatter(coord[0], coord[1])\n",
    "    plt.annotate(word, (coord[0], coord[1]))\n",
    "\n",
    "plt.savefig('img.jpg')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e116aadf",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.9.1"
  },
  "vscode": {
   "interpreter": {
    "hash": "f06e3119828c145c4028f0fee76387e74b42caa79bccb80b4ba64fe213bb9e06"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}