Ch 12. CNN - 이미지 분류를 응용한 문서 분류.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Chapter 12. CNN - 이미지 분류를 응용한 문서 분류\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. CNN을 이용한 문서 분류\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.1 NLTK 영화 리뷰 데이터 변환\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "from nltk.corpus import movie_reviews\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.preprocessing.text import Tokenizer\n",
    "from tensorflow.keras.preprocessing.sequence import pad_sequences\n",
    "from sklearn.model_selection import train_test_split \n",
    "\n",
    "#movie review data에서 file id를 가져옴\n",
    "fileids = movie_reviews.fileids() \n",
    "#file id를 이용해 raw text file을 가져옴\n",
    "reviews = [movie_reviews.raw(fileid) for fileid in fileids] \n",
    "categories = [movie_reviews.categories(fileid)[0] for fileid in fileids] \n",
    "\n",
    "np.random.seed(7)\n",
    "tf.random.set_seed(7)\n",
    "\n",
    "max_words = 10000 #사용할 단어의 수\n",
    "maxlen = 500 #문서의 단어 수를 제한\n",
    "\n",
    "# 빈도가 높은 10000개의 단어를 선택하도록 객체 생성\n",
    "tokenizer = Tokenizer(num_words=max_words, oov_token='UNK') \n",
    "tokenizer.fit_on_texts(reviews) #단어 인덱스 구축\n",
    "\n",
    "X = tokenizer.texts_to_sequences(reviews) #만들어진 단어 인덱스를 이용해 변환\n",
    "X = pad_sequences(X, maxlen=maxlen, truncating='pre') \n",
    "\n",
    "# label을 0, 1의 값으로 변환\n",
    "label_dict = {'pos':0, 'neg':1}\n",
    "y = np.array([label_dict[c] for c in categories])\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2.2 CNN 모형 생성 및 학습\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding (Embedding)        (None, 500, 64)           640000    \n",
      "_________________________________________________________________\n",
      "conv1d (Conv1D)              (None, 496, 128)          41088     \n",
      "_________________________________________________________________\n",
      "max_pooling1d (MaxPooling1D) (None, 248, 128)          0         \n",
      "_________________________________________________________________\n",
      "conv1d_1 (Conv1D)            (None, 244, 256)          164096    \n",
      "_________________________________________________________________\n",
      "max_pooling1d_1 (MaxPooling1 (None, 122, 256)          0         \n",
      "_________________________________________________________________\n",
      "flatten (Flatten)            (None, 31232)             0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 64)                1998912   \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 1)                 65        \n",
      "=================================================================\n",
      "Total params: 2,844,161\n",
      "Trainable params: 2,844,161\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense, Conv1D, MaxPooling1D\n",
    "from tensorflow.keras.layers import Embedding, Dropout, Flatten\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "\n",
    "model = Sequential([\n",
    "    #word embedding layer 생성\n",
    "    Embedding(max_words, 64, input_length=maxlen), \n",
    "    Conv1D(128, # 채널의 수\n",
    "            5, # 1D 필터 크기\n",
    "            padding='valid', \n",
    "            activation='relu',\n",
    "            strides=1), \n",
    "    MaxPooling1D(),\n",
    "    Conv1D(256, # 채널의 수\n",
    "            5, # 1D 필터 크기\n",
    "            padding='valid', \n",
    "            activation='relu',\n",
    "            strides=1), \n",
    "    MaxPooling1D(),\n",
    "    Flatten(),\n",
    "    Dense(64, activation='relu'),\n",
    "    Dense(1, activation='sigmoid') #binary logistic regression을 수행\n",
    "])\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "adam = Adam(learning_rate=1e-3)\n",
    "model.compile(optimizer=adam, loss='binary_crossentropy', metrics=['acc'])\n",
    "history = model.fit(X_train, y_train, \n",
    "                    epochs=20,\n",
    "                    batch_size=256,\n",
    "                    verbose=0,\n",
    "                    validation_split=0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "13/13 [==============================] - 0s 17ms/step - loss: 0.6965 - acc: 0.8175\n",
      "#Test accuracy:0.817\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "def plot_results(history, metric):\n",
    "    plt.plot(history.history[metric], 'b', label='Training '+metric)\n",
    "    plt.plot(history.history['val_'+metric], 'r--', label='Validation '+metric)\n",
    "    plt.title('Training vs. Validation '+metric)\n",
    "    plt.xlabel('Epochs')\n",
    "    plt.ylabel(metric)\n",
    "    plt.legend()\n",
    "    plt.show()\n",
    "\n",
    "plot_results(history, 'loss')\n",
    "\n",
    "#테스트 셋으로 학습된 모형의 성능을 평가\n",
    "score = model.evaluate(X_test, y_test)\n",
    "print(f'#Test accuracy:{score[1]:.3f}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "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": "3.8.0"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": true,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": true
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}