0402-klue-dnn-output.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "M4_PHrOVj51Z"
   },
   "source": [
    "🌱 인프런 📚 모두의 한국어 텍스트 분석과 자연어처리 with 파이썬 🐍 https://inf.run/FX4TP\n",
    "\n",
    "\n",
    "[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/corazzon/python-text-analysis/blob/main/0402-klue-dnn-output.ipynb)\n",
    "\n",
    "\n",
    "## DNN (Deep Neural Network) 으로 텍스트 분류하기\n",
    "\n",
    "\n",
    "## 라이브러리 로드"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "id": "4CcTgpxw6Aai"
   },
   "outputs": [],
   "source": [
    "import re\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import warnings\n",
    "\n",
    "# 경고 메시지 무시\n",
    "warnings.filterwarnings(\"ignore\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "hbUdn24iZ0JI"
   },
   "outputs": [],
   "source": [
    "# tensorflow\n",
    "# TextVectorization, Embedding, GlobalAveragePooling1D\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras import Sequential\n",
    "from tensorflow.keras.layers import TextVectorization, Embedding\n",
    "from tensorflow.keras.layers import GlobalAvgPool1D, Dense, Dropout"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ESivktS1lEm7"
   },
   "source": [
    "## 데이터 미리보기 및 요약"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 424
    },
    "id": "eFKC0rSZqNsO",
    "outputId": "fe1d906f-9cb4-4da6-da3b-cbf5c91e278d"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((45654, 4), (9131, 3))"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 학습, 예측 데이터셋을 불러옵니다.\n",
    "# pecab 으로 조사, 구두점을 제거한 데이터를 불러옵니다. \n",
    "# 전처리 되지 않은 텍스트를 사용해도 상관 없습니다.\n",
    "train = pd.read_csv(\"data/klue/train_pecab.csv\")\n",
    "test = pd.read_csv(\"data/klue/test_pecab.csv\")\n",
    "train.shape, test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "id": "TjuWJzWqZ0JF",
    "outputId": "26679284-ba4f-424b-c17b-41f8a5aad382"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "topic_idx\n",
       "4    0.167105\n",
       "2    0.161256\n",
       "5    0.151860\n",
       "6    0.147873\n",
       "1    0.136286\n",
       "3    0.129956\n",
       "0    0.105664\n",
       "Name: proportion, dtype: float64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# value_counts()로 분류별 빈도수 확인합니다.\n",
    "train[\"topic_idx\"].value_counts(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "id": "zvpvDX5nZ0JG"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'topic_idx'"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 정답(label) 값을 설정합니다.\n",
    "label_name = \"topic_idx\"\n",
    "label_name"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "sVldgQfrZ0JH"
   },
   "source": [
    "## label one-hot 형태로 만들기"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(45654, 7)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# from keras.utils import to_categorical\n",
    "from keras.utils import to_categorical\n",
    "# get_dummies 를 사용하여 label 값을 one-hot 형태로 만듭니다.\n",
    "y = to_categorical(train[label_name])\n",
    "y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "topic_idx\n",
       "0    4824\n",
       "1    6222\n",
       "2    7362\n",
       "3    5933\n",
       "4    7629\n",
       "5    6933\n",
       "6    6751\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# value_counts 로 정답 빈도 보기\n",
    "train[label_name].value_counts().sort_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# train[label_name]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    4824.0\n",
       "1    6222.0\n",
       "2    7362.0\n",
       "3    5933.0\n",
       "4    7629.0\n",
       "5    6933.0\n",
       "6    6751.0\n",
       "dtype: float32"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# one-hot-encoding 정답 빈도 보기\n",
    "pd.DataFrame(y).sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 검증용 데이터 만들기와 데이터 누수\n",
    "\n",
    "* 아래의 경우 데이터 누수(Data Leakage)에 해당됩니다.\n",
    "* label encoding, one-hot encoding 시 test 데이터 셋 활용하여 encoder를 fit하는 경우\n",
    "* data scaling 적용 시 test 데이터 셋 활용하여 scaler를 fit하는 경우\n",
    "* pandas의 get_dummies() 함수를 test 데이터셋에 적용하는 경우\n",
    "* test 데이터 셋의 결측치 처리 시 test 데이터 셋의 통계 값 활용\n",
    "* test 데이터 셋을 EDA하여 얻은 인사이트를 통해 학습에 활용하는 경우\n",
    "* test 데이터 셋을 학습 과정에 사용하는 모든 행위 (test 데이터셋은 추론에만 활용되어야 합니다)\n",
    "* test 데이터 셋의 데이터 개수 정보를 활용하는 경우 (실제 test 데이터셋은 몇개가 입력으로 들어올 지 모르기 때문)\n",
    "* 위 예시 외에도 test 데이터 셋이 모델 학습에 활용되는 경우에 Data leakage에 해당됨.\n",
    "* 출처 : https://dacon.io/competitions/official/236055/talkboard/407731 \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = train[\"title\"]\n",
    "X_test_text = test[\"title\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((36523,), (9131,), (36523, 7), (9131, 7))"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# train_test_split\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "X_train_text, X_valid_text, y_train, y_valid = train_test_split(\n",
    "    X, y, test_size=0.2, random_state=42, stratify=y)\n",
    "X_train_text.shape, X_valid_text.shape, y_train.shape, y_valid.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    0.105659\n",
       "1    0.136298\n",
       "2    0.161268\n",
       "3    0.129946\n",
       "4    0.167100\n",
       "5    0.151850\n",
       "6    0.147879\n",
       "dtype: float32"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 층화표집이 잘 되었는지 확인\n",
    "pd.DataFrame(y_train).mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    0.105684\n",
       "1    0.136239\n",
       "2    0.161209\n",
       "3    0.129997\n",
       "4    0.167123\n",
       "5    0.151900\n",
       "6    0.147848\n",
       "dtype: float32"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(y_valid).mean()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "5Vm0_ExvnAyN"
   },
   "source": [
    "## TextVectorization\n",
    "\n",
    "* 매개변수\n",
    "    * max_tokens: 텍스트에서 고려할 최대 토큰 수를 지정합니다. 가장 빈도가 높은 max_tokens 개의 토큰만 사용됩니다.\n",
    "    * standardize: 텍스트를 표준화하는 방법을 선택합니다. 'lower_and_strip_punctuation'로 설정하면 소문자로 변환하고 문장부호를 제거합니다.\n",
    "    * split: 텍스트를 토큰화하는 방법을 선택합니다. 'whitespace'로 설정하면 공백을 기준으로 텍스트가 토큰화됩니다.\n",
    "    * ngrams: 단일 토큰 이상의 n-그램을 생성하도록 지정합니다. 예를 들어, ngrams=2로 설정하면 바이그램이 생성됩니다.\n",
    "    * output_mode: 출력 모드를 지정합니다. 'int'로 설정하면 정수로 변환된 토큰이 출력됩니다.\n",
    "    * output_sequence_length: 출력 시퀀스의 길이를 지정합니다. 이 매개변수를 사용하여 모든 시퀀스를 고정된 길이로 패딩하거나 자를 수 있습니다.\n",
    "    * pad_to_max_tokens: True로 설정하면 max_tokens으로 시퀀스를 패딩합니다.\n",
    "    * vocabulary: 사용할 어휘 사전을 지정합니다. 사전이 제공되면 모델은 이 사전에 있는 토큰만 고려합니다.\n",
    "    * sparse: True로 설정하면 희소 텐서를 출력합니다.\n",
    "    * ragged: True로 설정하면 불규칙한 길이의 토큰을 나타내는 tf.RaggedTensor를 출력합니다.\n",
    "    * encoding: 텍스트의 인코딩을 지정합니다.\n",
    "    * kwargs: 기타 설정을 위한 추가적인 키워드 인수를 받습니다.\n",
    "\n",
    "\n",
    "* API Document: https://www.tensorflow.org/api_docs/python/tf/keras/preprocessing/text/Tokenizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'!\"#$%&\\'()*+,-./:;<=>?@[\\\\]^_`{|}~'"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import string\n",
    "string.punctuation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 소문자 변환: 입력된 텍스트를 모두 소문자로 변경합니다.\n",
    "# HTML 태그 제거: HTML 태그(<br />)를 공백으로 대체합니다.\n",
    "# 구두점 제거: 모든 구두점을 제거합니다.\n",
    "\n",
    "def custom_standardization(input_data):\n",
    "    lowercase = tf.strings.lower(input_data)\n",
    "    stripped_html = tf.strings.regex_replace(lowercase, '<br />', ' ')\n",
    "    return tf.strings.regex_replace(stripped_html,\n",
    "                                  '[%s]' % re.escape(string.punctuation), '')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## MAX_SEQUENCE_LENGTH \n",
    "* 단어 수에 대한 빈도 수 고려하기"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    45654.000000\n",
       "mean         6.588207\n",
       "std          1.474083\n",
       "min          1.000000\n",
       "25%          6.000000\n",
       "50%          7.000000\n",
       "75%          8.000000\n",
       "max         13.000000\n",
       "Name: title, dtype: float64"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 단어 빈도수의 기술 통계 구하기\n",
    "train[\"title\"].map(lambda x: len(x.split())).describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# TextVectorization 레이어 설정\n",
    "# standardize, max_tokens, output_mode\n",
    "vocab_size = 10000\n",
    "MAX_SEQUENCE_LENGTH = 10\n",
    "\n",
    "vectorizer = TextVectorization(standardize=custom_standardization, \n",
    "                               max_tokens=vocab_size, \n",
    "                               output_sequence_length=MAX_SEQUENCE_LENGTH)\n",
    "# X_train_text 만 넣어주면 valid 에 있는 어휘는 학습이 되지 않습니다.\n",
    "# test 에만 있는 어휘는 데이터 누수 문제로 사용하지 않습니다.\n",
    "vectorizer.adapt(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(45654,)"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tf.Tensor: shape=(36523, 10), dtype=int64, numpy=\n",
       "array([[   4,    1, 1391, ...,    0,    0,    0],\n",
       "       [ 531,  102,    1, ...,  437,    0,    0],\n",
       "       [2457, 7737,    1, ...,    0,    0,    0],\n",
       "       ...,\n",
       "       [   1,    1,   81, ...,    0,    0,    0],\n",
       "       [ 105, 2300,   84, ...,    0,    0,    0],\n",
       "       [ 898, 9971,    1, ...,  554,    0,    0]])>"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train = vectorizer(X_train_text)\n",
    "X_train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tf.Tensor: shape=(9131, 10), dtype=int64, numpy=\n",
       "array([[  23, 1924,  569, ...,    0,    0,    0],\n",
       "       [   1,    1,    1, ...,    0,    0,    0],\n",
       "       [ 130,   65,    1, ...,    0,    0,    0],\n",
       "       ...,\n",
       "       [   1,  171,  615, ..., 3147,    0,    0],\n",
       "       [   1,    1, 1930, ...,    0,    0,    0],\n",
       "       [3609, 1861, 5525, ...,    0,    0,    0]])>"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_valid = vectorizer(X_valid_text)\n",
    "X_valid"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tf.Tensor: shape=(9131, 10), dtype=int64, numpy=\n",
       "array([[1179,   88,    1, ...,    0,    0,    0],\n",
       "       [   1, 3462,    1, ...,    0,    0,    0],\n",
       "       [1731,    1,  940, ...,    0,    0,    0],\n",
       "       ...,\n",
       "       [3265,   19,    1, ...,  278,   36,    0],\n",
       "       [   4, 5895,    1, ...,    0,    0,    0],\n",
       "       [1503, 8864,    1, ...,    0,    0,    0]])>"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test = vectorizer(X_test_text)\n",
    "X_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10000, ['', '[UNK]', '첫', '등', '게시판', '北', '신간', '대통령', '감독', '美'])"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vocab = vectorizer.get_vocabulary()\n",
    "len(vocab), vocab[:10]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 텍스트 길이 시각화"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: >"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df_train = pd.DataFrame(X_train)\n",
    "sns.heatmap(df_train.replace(0, np.nan).notnull(), cmap=\"gray_r\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###  vocab index 분포\n",
    "* 0에 가까울 수록 빈도수가 높음"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: >"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# vocab index 분포 : 0에 가까울 수록 빈도수가 높음\n",
    "\n",
    "sns.heatmap(df_train, cmap=\"Blues\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 임베딩 \n",
    "\n",
    "* https://www.tensorflow.org/text/guide/word_embeddings?hl=ko\n",
    "* https://damien0x0023.github.io/rnnExplainer/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "id": "MbtT1Vdebei7"
   },
   "outputs": [],
   "source": [
    "# 하이퍼파라미터(모델링할 때 사용자가 직접 세팅해주는 값)을 설정합니다.\n",
    "# vocab_size는 텍스트 데이터의 전체 단어 집합의 크기입니다.\n",
    "# embedding_dim는 임베딩 할 벡터의 차원입니다.\n",
    "# max_length는 패딩의 기준이 됩니다.\n",
    "\n",
    "embedding_dim = 64"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(TensorShape([36523, 10]), 64)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape, embedding_dim"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "TensorShape([36523, 10, 64])"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "embedding_layer = Embedding(input_dim=vocab_size, \n",
    "                            output_dim=embedding_dim, \n",
    "                            input_length=MAX_SEQUENCE_LENGTH)\n",
    "embedding_train = embedding_layer(X_train)\n",
    "embedding_train.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## GlobalAvgPool1D\n",
    "\n",
    "* 고정된 출력 크기: GlobalAvgPool1D를 사용하면 입력 텍스트의 길이에 관계없이 고정된 출력 크기를 얻을 수 있습니다. 이는 다양한 길이의 문장을 처리하고 동일한 크기의 임베딩을 얻을 수 있도록 해줍니다.\n",
    "* 전역 정보 캡처: GlobalAvgPool1D는 각 시퀀스 차원에 대한 평균을 계산합니다. 이로써 각 차원의 전역 정보를 캡처할 수 있습니다. 이는 문장의 중요한 특징을 추출하는 데 도움이 되며, 특히 문장이 길거나 다양한 길이일 때 유용합니다.\n",
    "* 과적합 방지: GlobalAvgPool1D를 사용하면 모델이 입력의 특정 위치에 과도하게 의존하는 것을 방지할 수 있습니다. 이는 모델을 더욱 일반화되게 만들어주어 새로운 데이터에 대한 성능을 향상시킬 수 있습니다.\n",
    "* 계산 효율성: GlobalAvgPool1D는 입력의 평균을 계산하므로 모델이 더 적은 파라미터로 텍스트를 처리할 수 있습니다. 이는 모델의 효율성을 높여줍니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: >"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# avg_pooling_layer heatmap\n",
    "avg_pooling_layer = GlobalAvgPool1D()\n",
    "sns.heatmap(avg_pooling_layer(embedding_train), cmap=\"Blues\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "uInDm3wobb5v"
   },
   "source": [
    "## 층 구성\n",
    " <img src=\"https://cs231n.github.io/assets/nn1/neural_net2.jpeg\" width=\"40%\">\n",
    "\n",
    "* 이미지 출처 : https://cs231n.github.io/neural-networks-1/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "id": "ATWh1mLDZ0JN",
    "outputId": "a30b1258-17b2-4b5b-e36e-ecc3e269b8eb"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "7"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 클래스의 수는 분류될 예측값의 종류입니다.\n",
    "# 정답값이 one-hot 형태로 인코딩 되어 있기 때문에 정답값의 컬럼의 수가 예측값의 종류가 됩니다.\n",
    "n_class = y.shape[1]\n",
    "n_class"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "id": "VyRptbe9bjP-"
   },
   "outputs": [],
   "source": [
    "# Sequential Layer\n",
    "\n",
    "model = Sequential()\n",
    "model.add(embedding_layer)\n",
    "model.add(avg_pooling_layer)\n",
    "model.add(Dense(units=32))\n",
    "model.add(Dense(units=32))\n",
    "model.add(Dropout(0.1))\n",
    "model.add(Dense(units=n_class, activation=\"softmax\"))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "TD6brH6LZ0JN"
   },
   "source": [
    "## 모델 컴파일"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "id": "_uYPV_gbZ0JO",
    "outputId": "0240ab38-6945-433a-e32e-d383d5ddfbd6"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " embedding (Embedding)       (None, 10, 64)            640000    \n",
      "                                                                 \n",
      " global_average_pooling1d (  (None, 64)                0         \n",
      " GlobalAveragePooling1D)                                         \n",
      "                                                                 \n",
      " dense (Dense)               (None, 32)                2080      \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 32)                1056      \n",
      "                                                                 \n",
      " dropout (Dropout)           (None, 32)                0         \n",
      "                                                                 \n",
      " dense_2 (Dense)             (None, 7)                 231       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 643367 (2.45 MB)\n",
      "Trainable params: 643367 (2.45 MB)\n",
      "Non-trainable params: 0 (0.00 Byte)\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# 여러개 정답 중 하나 맞추는 문제이며, 정답값이 one-hot 형태이기 때문에\n",
    "# 손실 함수는 categorical_crossentropy를 사용합니다.\n",
    "model.compile(loss=\"categorical_crossentropy\", \n",
    "              optimizer=tf.keras.optimizers.Adam(learning_rate=0.0001), \n",
    "              metrics=[\"accuracy\"]\n",
    "             )\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "K6Z8ZNu-Z0JO"
   },
   "source": [
    "## 학습 fit\n",
    "\n",
    "<img src=\"https://i.imgur.com/uWkUCu1.png\" width=\"700\">\n",
    "\n",
    "* https://www.tensorflow.org/tensorboard/get_started?hl=ko"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Clear any logs from previous runs\n",
    "# !rm -rf ./logs/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.callbacks import EarlyStopping, TensorBoard\n",
    "from datetime import datetime\n",
    "\n",
    "log_dir = \"logs/fit/\" + datetime.now().strftime(\"%Y%m%d-%H%M%S\")\n",
    "\n",
    "tensorboard_callback = TensorBoard(log_dir=log_dir)\n",
    "early_stop = EarlyStopping(monitor='val_loss', patience=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "o-IIagNa8n9z",
    "outputId": "f724cbb3-c7ef-4df6-c14c-e8f324132f30",
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "571/571 [==============================] - 8s 12ms/step - loss: 1.9126 - accuracy: 0.2497 - val_loss: 1.8528 - val_accuracy: 0.3536\n",
      "Epoch 2/100\n",
      "571/571 [==============================] - 8s 14ms/step - loss: 1.6781 - accuracy: 0.4667 - val_loss: 1.4694 - val_accuracy: 0.5604\n",
      "Epoch 3/100\n",
      "571/571 [==============================] - 9s 16ms/step - loss: 1.2634 - accuracy: 0.6219 - val_loss: 1.1327 - val_accuracy: 0.6495\n",
      "Epoch 4/100\n",
      "571/571 [==============================] - 12s 20ms/step - loss: 0.9735 - accuracy: 0.7109 - val_loss: 0.9412 - val_accuracy: 0.7080\n",
      "Epoch 5/100\n",
      "571/571 [==============================] - 12s 21ms/step - loss: 0.8017 - accuracy: 0.7573 - val_loss: 0.8483 - val_accuracy: 0.7271\n",
      "Epoch 6/100\n",
      "571/571 [==============================] - 12s 21ms/step - loss: 0.7023 - accuracy: 0.7856 - val_loss: 0.8040 - val_accuracy: 0.7407\n",
      "Epoch 7/100\n",
      "571/571 [==============================] - 13s 23ms/step - loss: 0.6335 - accuracy: 0.8055 - val_loss: 0.7795 - val_accuracy: 0.7503\n",
      "Epoch 8/100\n",
      "571/571 [==============================] - 14s 24ms/step - loss: 0.5815 - accuracy: 0.8212 - val_loss: 0.7635 - val_accuracy: 0.7563\n",
      "Epoch 9/100\n",
      "571/571 [==============================] - 14s 24ms/step - loss: 0.5389 - accuracy: 0.8319 - val_loss: 0.7554 - val_accuracy: 0.7629\n",
      "Epoch 10/100\n",
      "571/571 [==============================] - 14s 25ms/step - loss: 0.5009 - accuracy: 0.8436 - val_loss: 0.7495 - val_accuracy: 0.7683\n",
      "Epoch 11/100\n",
      "571/571 [==============================] - 13s 23ms/step - loss: 0.4680 - accuracy: 0.8541 - val_loss: 0.7479 - val_accuracy: 0.7715\n",
      "Epoch 12/100\n",
      "571/571 [==============================] - 14s 25ms/step - loss: 0.4370 - accuracy: 0.8638 - val_loss: 0.7509 - val_accuracy: 0.7717\n",
      "Epoch 13/100\n",
      "571/571 [==============================] - 14s 24ms/step - loss: 0.4118 - accuracy: 0.8710 - val_loss: 0.7547 - val_accuracy: 0.7707\n",
      "Epoch 14/100\n",
      "571/571 [==============================] - 15s 27ms/step - loss: 0.3894 - accuracy: 0.8764 - val_loss: 0.7621 - val_accuracy: 0.7723\n",
      "Epoch 15/100\n",
      "571/571 [==============================] - 13s 23ms/step - loss: 0.3679 - accuracy: 0.8829 - val_loss: 0.7733 - val_accuracy: 0.7723\n",
      "Epoch 16/100\n",
      "571/571 [==============================] - 12s 22ms/step - loss: 0.3508 - accuracy: 0.8879 - val_loss: 0.7884 - val_accuracy: 0.7723\n"
     ]
    }
   ],
   "source": [
    "# 모델 학습을 실행합니다.\n",
    "history = model.fit(X_train, y_train, \n",
    "                    epochs=100, batch_size=64,\n",
    "                    callbacks=[early_stop, tensorboard_callback],\n",
    "                    validation_data=(X_valid, y_valid), \n",
    "                    use_multiprocessing=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "id": "LMnumSxuZ0JO",
    "outputId": "5e243138-3f6d-43bd-c220-0d17762e4cdd"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<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>loss</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>val_loss</th>\n",
       "      <th>val_accuracy</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.912610</td>\n",
       "      <td>0.249706</td>\n",
       "      <td>1.852776</td>\n",
       "      <td>0.353630</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.678145</td>\n",
       "      <td>0.466747</td>\n",
       "      <td>1.469357</td>\n",
       "      <td>0.560399</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.263371</td>\n",
       "      <td>0.621909</td>\n",
       "      <td>1.132696</td>\n",
       "      <td>0.649545</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.973474</td>\n",
       "      <td>0.710922</td>\n",
       "      <td>0.941204</td>\n",
       "      <td>0.708028</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.801713</td>\n",
       "      <td>0.757331</td>\n",
       "      <td>0.848296</td>\n",
       "      <td>0.727084</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.702311</td>\n",
       "      <td>0.785642</td>\n",
       "      <td>0.804029</td>\n",
       "      <td>0.740664</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.633533</td>\n",
       "      <td>0.805492</td>\n",
       "      <td>0.779507</td>\n",
       "      <td>0.750301</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.581490</td>\n",
       "      <td>0.821181</td>\n",
       "      <td>0.763532</td>\n",
       "      <td>0.756325</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.538926</td>\n",
       "      <td>0.831887</td>\n",
       "      <td>0.755379</td>\n",
       "      <td>0.762896</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.500897</td>\n",
       "      <td>0.843633</td>\n",
       "      <td>0.749486</td>\n",
       "      <td>0.768262</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.468015</td>\n",
       "      <td>0.854119</td>\n",
       "      <td>0.747853</td>\n",
       "      <td>0.771547</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.436965</td>\n",
       "      <td>0.863784</td>\n",
       "      <td>0.750858</td>\n",
       "      <td>0.771657</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.411826</td>\n",
       "      <td>0.870958</td>\n",
       "      <td>0.754743</td>\n",
       "      <td>0.770671</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.389436</td>\n",
       "      <td>0.876407</td>\n",
       "      <td>0.762060</td>\n",
       "      <td>0.772314</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.367918</td>\n",
       "      <td>0.882923</td>\n",
       "      <td>0.773295</td>\n",
       "      <td>0.772314</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.350771</td>\n",
       "      <td>0.887851</td>\n",
       "      <td>0.788391</td>\n",
       "      <td>0.772314</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        loss  accuracy  val_loss  val_accuracy\n",
       "0   1.912610  0.249706  1.852776      0.353630\n",
       "1   1.678145  0.466747  1.469357      0.560399\n",
       "2   1.263371  0.621909  1.132696      0.649545\n",
       "3   0.973474  0.710922  0.941204      0.708028\n",
       "4   0.801713  0.757331  0.848296      0.727084\n",
       "5   0.702311  0.785642  0.804029      0.740664\n",
       "6   0.633533  0.805492  0.779507      0.750301\n",
       "7   0.581490  0.821181  0.763532      0.756325\n",
       "8   0.538926  0.831887  0.755379      0.762896\n",
       "9   0.500897  0.843633  0.749486      0.768262\n",
       "10  0.468015  0.854119  0.747853      0.771547\n",
       "11  0.436965  0.863784  0.750858      0.771657\n",
       "12  0.411826  0.870958  0.754743      0.770671\n",
       "13  0.389436  0.876407  0.762060      0.772314\n",
       "14  0.367918  0.882923  0.773295      0.772314\n",
       "15  0.350771  0.887851  0.788391      0.772314"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 모델 학습의 결과값을 데이터 프레임으로 만들어 확인합니다.\n",
    "df_hist = pd.DataFrame(history.history)\n",
    "df_hist"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 315
    },
    "id": "MCU7v3_Ec5eA",
    "outputId": "85ad719d-b5a7-4f8f-a34d-734b40f2b427"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: >"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 모델 학습 결과을 그래프로 시각화합니다.\n",
    "df_hist[[\"accuracy\", \"val_accuracy\"]].plot()\n",
    "\n",
    "df_hist[[\"loss\", \"val_loss\"]].plot()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## tensorboard"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "# !kill 12560"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "      <iframe id=\"tensorboard-frame-7010641da6e6dee7\" width=\"100%\" height=\"800\" frameborder=\"0\">\n",
       "      </iframe>\n",
       "      <script>\n",
       "        (function() {\n",
       "          const frame = document.getElementById(\"tensorboard-frame-7010641da6e6dee7\");\n",
       "          const url = new URL(\"/\", window.location);\n",
       "          const port = 6006;\n",
       "          if (port) {\n",
       "            url.port = port;\n",
       "          }\n",
       "          frame.src = url;\n",
       "        })();\n",
       "      </script>\n",
       "    "
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# docs_infra: no_execute\n",
    "%load_ext tensorboard\n",
    "# %reload_ext tensorboard\n",
    "%tensorboard --logdir logs"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "OFpNLGzUZ0JP"
   },
   "source": [
    "## 예측"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "o-jaBsHTt7Ri",
    "outputId": "e3e24efa-b451-4e2d-a594-953acadd4231"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "286/286 [==============================] - 1s 2ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([[2.4454135e-01, 4.1179629e-03, 1.5380041e-01, ..., 1.2828516e-04,\n",
       "        1.2370635e-05, 1.6778093e-03],\n",
       "       [1.1874215e-03, 4.4177400e-06, 3.4808301e-04, ..., 7.2191765e-06,\n",
       "        3.2851528e-08, 8.9278049e-04],\n",
       "       [2.4073562e-03, 2.1918350e-01, 3.5345295e-01, ..., 3.9968488e-01,\n",
       "        2.2719473e-02, 1.1027746e-03],\n",
       "       ...,\n",
       "       [2.2170539e-03, 2.9119133e-04, 6.8074125e-01, ..., 1.5858679e-01,\n",
       "        8.2471361e-06, 1.1581543e-04],\n",
       "       [6.4938427e-03, 1.2293403e-03, 9.5124614e-01, ..., 5.3470023e-04,\n",
       "        3.4290348e-04, 9.3817366e-03],\n",
       "       [1.0900661e-01, 3.5216730e-02, 4.9985304e-01, ..., 8.1781335e-03,\n",
       "        1.9215032e-02, 1.7389488e-01]], dtype=float32)"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# predict() 메서드로 모델 예측합니다.\n",
    "y_pred = model.predict(X_test)\n",
    "y_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "TensorShape([9131, 10])"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(9131, 7)"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.245, 0.004, 0.154, 0.596, 0.   , 0.   , 0.002], dtype=float32)"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred[0].round(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.argmax(y_pred[0].round(3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([3, 3, 4, ..., 2, 2, 2])"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# numpy.argmax를 이용해 가장 큰 값의 인덱스들을 반환한 값(클래스 예측)을 y_predict에 할당합니다.\n",
    "y_predict = np.argmax(y_pred, axis=1)\n",
    "y_predict"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "CTE70dO3Z0JP"
   },
   "source": [
    "## 평가\n",
    "* 실제 미래의 데이터나 경진대회 데이터는 test 의 정답을 모릅니다. \n",
    "* 그렇기 때문에 validation 점수를 평가 점수로 활용합니다."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 제출하기"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<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>index</th>\n",
       "      <th>topic_idx</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>45654</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>45655</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>45656</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>45657</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>45658</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   index  topic_idx\n",
       "0  45654          0\n",
       "1  45655          0\n",
       "2  45656          0\n",
       "3  45657          0\n",
       "4  45658          0"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# sample_submission 파일을 불러옵니다.\n",
    "# Dacon 에 제출할 답안지 양식입니다.\n",
    "submit = pd.read_csv(\"data/klue/sample_submission.csv\")\n",
    "submit.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<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>index</th>\n",
       "      <th>topic_idx</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>45654</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>45655</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>45656</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>45657</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>45658</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9126</th>\n",
       "      <td>54780</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9127</th>\n",
       "      <td>54781</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9128</th>\n",
       "      <td>54782</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9129</th>\n",
       "      <td>54783</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9130</th>\n",
       "      <td>54784</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>9131 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      index  topic_idx\n",
       "0     45654          3\n",
       "1     45655          3\n",
       "2     45656          4\n",
       "3     45657          2\n",
       "4     45658          3\n",
       "...     ...        ...\n",
       "9126  54780          2\n",
       "9127  54781          6\n",
       "9128  54782          2\n",
       "9129  54783          2\n",
       "9130  54784          2\n",
       "\n",
       "[9131 rows x 2 columns]"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 정답값 측정을 위해 y_test 변수에 할당\n",
    "submit[\"topic_idx\"] = y_predict\n",
    "submit"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7723140716552734"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "valid_accuracy = df_hist[\"val_accuracy\"].iloc[-1]\n",
    "valid_accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'data/klue/submit_0.77231.csv'"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "file_name = f\"data/klue/submit_{valid_accuracy:.5f}.csv\"\n",
    "file_name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 잘 저장이 되었는지 확인합니다.\n",
    "submit.to_csv(file_name, index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<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>index</th>\n",
       "      <th>topic_idx</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>45654</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>45655</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>45656</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>45657</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>45658</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9126</th>\n",
       "      <td>54780</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9127</th>\n",
       "      <td>54781</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9128</th>\n",
       "      <td>54782</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9129</th>\n",
       "      <td>54783</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9130</th>\n",
       "      <td>54784</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>9131 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      index  topic_idx\n",
       "0     45654          3\n",
       "1     45655          3\n",
       "2     45656          4\n",
       "3     45657          2\n",
       "4     45658          3\n",
       "...     ...        ...\n",
       "9126  54780          2\n",
       "9127  54781          6\n",
       "9128  54782          2\n",
       "9129  54783          2\n",
       "9130  54784          2\n",
       "\n",
       "[9131 rows x 2 columns]"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 잘 저장이 되었는지 확인합니다.\n",
    "pd.read_csv(file_name)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* dacon에 제출해서 리더보드 확인하기 : https://dacon.io/competitions/official/235747/mysubmission\n",
    "\n",
    "\n",
    "## 여러 방법을 통해 예측비율을 높여보세요. \n",
    "* 아래 항목 외의 기법을 사용해도 됩니다.\n",
    " * 전처리 하기\n",
    " * 불용어 처리\n",
    " * BOW, TF-IDF의 파라메터 변경\n",
    " * 분류기의 파라메터 변경\n",
    " * 분류기 변경"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 임베딩 프로젝터 사용하기\n",
    "* 임베딩 프로젝터 : [링크](https://projector.tensorflow.org/?hl=ko&_gl=1*qcvijm*_ga*ODk4NzAxOTgzLjE3MDExNjQyNzM.*_ga_W0YLR4190T*MTcwMTgzNzYzMy42LjEuMTcwMTgzODU2Mi4wLjAuMA..)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " embedding (Embedding)       (None, 10, 64)            640000    \n",
      "                                                                 \n",
      " global_average_pooling1d (  (None, 64)                0         \n",
      " GlobalAveragePooling1D)                                         \n",
      "                                                                 \n",
      " dense (Dense)               (None, 32)                2080      \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 32)                1056      \n",
      "                                                                 \n",
      " dropout (Dropout)           (None, 32)                0         \n",
      "                                                                 \n",
      " dense_2 (Dense)             (None, 7)                 231       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 643367 (2.45 MB)\n",
      "Trainable params: 643367 (2.45 MB)\n",
      "Non-trainable params: 0 (0.00 Byte)\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['', '[UNK]', '첫', '등', '게시판']"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# No such layer: embedding. Existing layers are: ['embedding_2', 등의 메시지가 나온다면\n",
    "# 컴파일을 여러번 해서 그렇습니다. embedding_2 등으로 변경해 주거나 해당 주피터 노트북의 커널 메뉴에서 재시작을 해주시면 됩니다.\n",
    "# weights\n",
    "# vocab\n",
    "weights = model.get_layer(\"embedding\").get_weights()[0]\n",
    "vocab[:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<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>[UNK]</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>첫</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>등</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>게시판</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>北</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>신간</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9993</th>\n",
       "      <td>유전</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9994</th>\n",
       "      <td>유적지서</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9995</th>\n",
       "      <td>유재학</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9996</th>\n",
       "      <td>유작</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9997</th>\n",
       "      <td>유임</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>9998 rows × 1 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     [UNK]\n",
       "0        첫\n",
       "1        등\n",
       "2      게시판\n",
       "3        北\n",
       "4       신간\n",
       "...    ...\n",
       "9993    유전\n",
       "9994  유적지서\n",
       "9995   유재학\n",
       "9996    유작\n",
       "9997    유임\n",
       "\n",
       "[9998 rows x 1 columns]"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 가장 첫 데이터는 패딩으로 이후 데이터 부터 사용\n",
    "# 파일로 저장 : \"metadata.tsv\",  sep=\"\\t\", index=False, header=False\n",
    "df_vocab = pd.DataFrame(vocab[1:])\n",
    "df_vocab.to_csv(\"metadata.tsv\", sep=\"\\t\", index=False, header=False)\n",
    "pd.read_csv(\"metadata.tsv\", sep=\"\\t\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "10000"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(vocab)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<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>-0.0030594869</th>\n",
       "      <th>-0.021908686</th>\n",
       "      <th>-0.006332541</th>\n",
       "      <th>-0.029790018</th>\n",
       "      <th>-0.014056583</th>\n",
       "      <th>0.019522417</th>\n",
       "      <th>-0.021364052</th>\n",
       "      <th>-0.026322097</th>\n",
       "      <th>0.013077043</th>\n",
       "      <th>0.048947953</th>\n",
       "      <th>...</th>\n",
       "      <th>-0.018743865</th>\n",
       "      <th>0.005321037</th>\n",
       "      <th>-0.0032379369</th>\n",
       "      <th>0.034092676</th>\n",
       "      <th>-0.02595583</th>\n",
       "      <th>0.037519284</th>\n",
       "      <th>0.033254053</th>\n",
       "      <th>0.043791447</th>\n",
       "      <th>-0.04261065</th>\n",
       "      <th>-0.05009249</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-0.060886</td>\n",
       "      <td>0.053186</td>\n",
       "      <td>0.035159</td>\n",
       "      <td>0.077219</td>\n",
       "      <td>0.027470</td>\n",
       "      <td>0.031742</td>\n",
       "      <td>-0.026134</td>\n",
       "      <td>0.059799</td>\n",
       "      <td>0.015636</td>\n",
       "      <td>-0.040177</td>\n",
       "      <td>...</td>\n",
       "      <td>0.027378</td>\n",
       "      <td>-0.015751</td>\n",
       "      <td>0.023973</td>\n",
       "      <td>0.014351</td>\n",
       "      <td>-0.005376</td>\n",
       "      <td>0.013189</td>\n",
       "      <td>-0.036368</td>\n",
       "      <td>0.001482</td>\n",
       "      <td>0.004892</td>\n",
       "      <td>-0.009380</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.049075</td>\n",
       "      <td>-0.078737</td>\n",
       "      <td>0.055392</td>\n",
       "      <td>-0.072645</td>\n",
       "      <td>-0.012353</td>\n",
       "      <td>-0.043689</td>\n",
       "      <td>0.035538</td>\n",
       "      <td>-0.064473</td>\n",
       "      <td>-0.002086</td>\n",
       "      <td>-0.001162</td>\n",
       "      <td>...</td>\n",
       "      <td>0.030059</td>\n",
       "      <td>0.001205</td>\n",
       "      <td>-0.053094</td>\n",
       "      <td>-0.031094</td>\n",
       "      <td>-0.007185</td>\n",
       "      <td>0.015966</td>\n",
       "      <td>0.042741</td>\n",
       "      <td>0.005806</td>\n",
       "      <td>0.036571</td>\n",
       "      <td>0.000183</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.068160</td>\n",
       "      <td>-0.199542</td>\n",
       "      <td>0.083718</td>\n",
       "      <td>0.011103</td>\n",
       "      <td>0.013463</td>\n",
       "      <td>-0.019877</td>\n",
       "      <td>0.037557</td>\n",
       "      <td>-0.131591</td>\n",
       "      <td>-0.149101</td>\n",
       "      <td>0.055411</td>\n",
       "      <td>...</td>\n",
       "      <td>0.004914</td>\n",
       "      <td>-0.009223</td>\n",
       "      <td>-0.192491</td>\n",
       "      <td>0.046734</td>\n",
       "      <td>0.059570</td>\n",
       "      <td>0.116772</td>\n",
       "      <td>-0.008699</td>\n",
       "      <td>0.037783</td>\n",
       "      <td>0.033965</td>\n",
       "      <td>0.007545</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.151854</td>\n",
       "      <td>0.343776</td>\n",
       "      <td>-0.010329</td>\n",
       "      <td>-0.062634</td>\n",
       "      <td>-0.048207</td>\n",
       "      <td>-0.084357</td>\n",
       "      <td>-0.025355</td>\n",
       "      <td>0.041701</td>\n",
       "      <td>0.123331</td>\n",
       "      <td>-0.140135</td>\n",
       "      <td>...</td>\n",
       "      <td>0.078058</td>\n",
       "      <td>-0.004534</td>\n",
       "      <td>-0.032620</td>\n",
       "      <td>0.038696</td>\n",
       "      <td>-0.025949</td>\n",
       "      <td>0.007402</td>\n",
       "      <td>0.058728</td>\n",
       "      <td>0.056710</td>\n",
       "      <td>0.010464</td>\n",
       "      <td>-0.045831</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.204413</td>\n",
       "      <td>-0.069902</td>\n",
       "      <td>-0.131237</td>\n",
       "      <td>0.322309</td>\n",
       "      <td>-0.227344</td>\n",
       "      <td>-0.141936</td>\n",
       "      <td>-0.141966</td>\n",
       "      <td>0.198913</td>\n",
       "      <td>-0.284505</td>\n",
       "      <td>0.208275</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.121993</td>\n",
       "      <td>0.225130</td>\n",
       "      <td>-0.165683</td>\n",
       "      <td>0.198171</td>\n",
       "      <td>0.236321</td>\n",
       "      <td>0.220491</td>\n",
       "      <td>-0.291206</td>\n",
       "      <td>0.062596</td>\n",
       "      <td>-0.038694</td>\n",
       "      <td>-0.291128</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9993</th>\n",
       "      <td>-0.023548</td>\n",
       "      <td>0.081437</td>\n",
       "      <td>-0.025308</td>\n",
       "      <td>0.012666</td>\n",
       "      <td>-0.058985</td>\n",
       "      <td>0.084137</td>\n",
       "      <td>0.063588</td>\n",
       "      <td>0.032992</td>\n",
       "      <td>-0.051704</td>\n",
       "      <td>0.077810</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.083701</td>\n",
       "      <td>-0.080253</td>\n",
       "      <td>0.073999</td>\n",
       "      <td>-0.023165</td>\n",
       "      <td>-0.090743</td>\n",
       "      <td>-0.069637</td>\n",
       "      <td>-0.056774</td>\n",
       "      <td>-0.004359</td>\n",
       "      <td>-0.051964</td>\n",
       "      <td>0.065627</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9994</th>\n",
       "      <td>0.016062</td>\n",
       "      <td>0.036560</td>\n",
       "      <td>-0.027451</td>\n",
       "      <td>0.073912</td>\n",
       "      <td>-0.031690</td>\n",
       "      <td>-0.001107</td>\n",
       "      <td>0.032623</td>\n",
       "      <td>0.095315</td>\n",
       "      <td>-0.033196</td>\n",
       "      <td>0.000488</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.056540</td>\n",
       "      <td>0.002616</td>\n",
       "      <td>0.030366</td>\n",
       "      <td>0.002743</td>\n",
       "      <td>-0.015085</td>\n",
       "      <td>0.027882</td>\n",
       "      <td>-0.053672</td>\n",
       "      <td>0.048045</td>\n",
       "      <td>-0.047991</td>\n",
       "      <td>-0.052953</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9995</th>\n",
       "      <td>-0.054924</td>\n",
       "      <td>0.061067</td>\n",
       "      <td>0.041327</td>\n",
       "      <td>-0.030925</td>\n",
       "      <td>0.042627</td>\n",
       "      <td>-0.014704</td>\n",
       "      <td>-0.063469</td>\n",
       "      <td>-0.049250</td>\n",
       "      <td>0.033490</td>\n",
       "      <td>0.009027</td>\n",
       "      <td>...</td>\n",
       "      <td>0.064704</td>\n",
       "      <td>-0.038800</td>\n",
       "      <td>0.003669</td>\n",
       "      <td>0.006371</td>\n",
       "      <td>-0.071749</td>\n",
       "      <td>-0.074866</td>\n",
       "      <td>-0.013044</td>\n",
       "      <td>-0.060729</td>\n",
       "      <td>-0.005727</td>\n",
       "      <td>0.058483</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9996</th>\n",
       "      <td>0.067798</td>\n",
       "      <td>-0.015144</td>\n",
       "      <td>-0.040362</td>\n",
       "      <td>0.086182</td>\n",
       "      <td>-0.003003</td>\n",
       "      <td>-0.031985</td>\n",
       "      <td>-0.010200</td>\n",
       "      <td>0.046585</td>\n",
       "      <td>-0.044793</td>\n",
       "      <td>0.004949</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.018300</td>\n",
       "      <td>0.011548</td>\n",
       "      <td>-0.002817</td>\n",
       "      <td>0.040713</td>\n",
       "      <td>0.000576</td>\n",
       "      <td>0.057323</td>\n",
       "      <td>-0.012639</td>\n",
       "      <td>-0.001651</td>\n",
       "      <td>-0.028798</td>\n",
       "      <td>-0.048197</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9997</th>\n",
       "      <td>0.054596</td>\n",
       "      <td>0.007665</td>\n",
       "      <td>0.053243</td>\n",
       "      <td>-0.000057</td>\n",
       "      <td>-0.026011</td>\n",
       "      <td>-0.002167</td>\n",
       "      <td>-0.011178</td>\n",
       "      <td>0.006316</td>\n",
       "      <td>0.047354</td>\n",
       "      <td>-0.024342</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.018951</td>\n",
       "      <td>0.020266</td>\n",
       "      <td>-0.064025</td>\n",
       "      <td>0.047593</td>\n",
       "      <td>0.010288</td>\n",
       "      <td>0.027306</td>\n",
       "      <td>0.065600</td>\n",
       "      <td>-0.048982</td>\n",
       "      <td>0.000822</td>\n",
       "      <td>-0.002681</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>9998 rows × 64 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      -0.0030594869  -0.021908686  -0.006332541  -0.029790018  -0.014056583  \\\n",
       "0         -0.060886      0.053186      0.035159      0.077219      0.027470   \n",
       "1          0.049075     -0.078737      0.055392     -0.072645     -0.012353   \n",
       "2          0.068160     -0.199542      0.083718      0.011103      0.013463   \n",
       "3          0.151854      0.343776     -0.010329     -0.062634     -0.048207   \n",
       "4          0.204413     -0.069902     -0.131237      0.322309     -0.227344   \n",
       "...             ...           ...           ...           ...           ...   \n",
       "9993      -0.023548      0.081437     -0.025308      0.012666     -0.058985   \n",
       "9994       0.016062      0.036560     -0.027451      0.073912     -0.031690   \n",
       "9995      -0.054924      0.061067      0.041327     -0.030925      0.042627   \n",
       "9996       0.067798     -0.015144     -0.040362      0.086182     -0.003003   \n",
       "9997       0.054596      0.007665      0.053243     -0.000057     -0.026011   \n",
       "\n",
       "      0.019522417  -0.021364052  -0.026322097  0.013077043  0.048947953  ...  \\\n",
       "0        0.031742     -0.026134      0.059799     0.015636    -0.040177  ...   \n",
       "1       -0.043689      0.035538     -0.064473    -0.002086    -0.001162  ...   \n",
       "2       -0.019877      0.037557     -0.131591    -0.149101     0.055411  ...   \n",
       "3       -0.084357     -0.025355      0.041701     0.123331    -0.140135  ...   \n",
       "4       -0.141936     -0.141966      0.198913    -0.284505     0.208275  ...   \n",
       "...           ...           ...           ...          ...          ...  ...   \n",
       "9993     0.084137      0.063588      0.032992    -0.051704     0.077810  ...   \n",
       "9994    -0.001107      0.032623      0.095315    -0.033196     0.000488  ...   \n",
       "9995    -0.014704     -0.063469     -0.049250     0.033490     0.009027  ...   \n",
       "9996    -0.031985     -0.010200      0.046585    -0.044793     0.004949  ...   \n",
       "9997    -0.002167     -0.011178      0.006316     0.047354    -0.024342  ...   \n",
       "\n",
       "      -0.018743865  0.005321037  -0.0032379369  0.034092676  -0.02595583  \\\n",
       "0         0.027378    -0.015751       0.023973     0.014351    -0.005376   \n",
       "1         0.030059     0.001205      -0.053094    -0.031094    -0.007185   \n",
       "2         0.004914    -0.009223      -0.192491     0.046734     0.059570   \n",
       "3         0.078058    -0.004534      -0.032620     0.038696    -0.025949   \n",
       "4        -0.121993     0.225130      -0.165683     0.198171     0.236321   \n",
       "...            ...          ...            ...          ...          ...   \n",
       "9993     -0.083701    -0.080253       0.073999    -0.023165    -0.090743   \n",
       "9994     -0.056540     0.002616       0.030366     0.002743    -0.015085   \n",
       "9995      0.064704    -0.038800       0.003669     0.006371    -0.071749   \n",
       "9996     -0.018300     0.011548      -0.002817     0.040713     0.000576   \n",
       "9997     -0.018951     0.020266      -0.064025     0.047593     0.010288   \n",
       "\n",
       "      0.037519284  0.033254053  0.043791447  -0.04261065  -0.05009249  \n",
       "0        0.013189    -0.036368     0.001482     0.004892    -0.009380  \n",
       "1        0.015966     0.042741     0.005806     0.036571     0.000183  \n",
       "2        0.116772    -0.008699     0.037783     0.033965     0.007545  \n",
       "3        0.007402     0.058728     0.056710     0.010464    -0.045831  \n",
       "4        0.220491    -0.291206     0.062596    -0.038694    -0.291128  \n",
       "...           ...          ...          ...          ...          ...  \n",
       "9993    -0.069637    -0.056774    -0.004359    -0.051964     0.065627  \n",
       "9994     0.027882    -0.053672     0.048045    -0.047991    -0.052953  \n",
       "9995    -0.074866    -0.013044    -0.060729    -0.005727     0.058483  \n",
       "9996     0.057323    -0.012639    -0.001651    -0.028798    -0.048197  \n",
       "9997     0.027306     0.065600    -0.048982     0.000822    -0.002681  \n",
       "\n",
       "[9998 rows x 64 columns]"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# \"vectors.tsv\"\n",
    "# index=False, header=False 로 저장\n",
    "df_w = pd.DataFrame(weights[1:len(vocab)])\n",
    "df_w.to_csv(\"vectors.tsv\", sep=\"\\t\", index=False, header=False)\n",
    "pd.read_csv(\"vectors.tsv\", sep=\"\\t\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* 임베딩 프로젝터 : [링크](https://projector.tensorflow.org/?hl=ko&_gl=1*qcvijm*_ga*ODk4NzAxOTgzLjE3MDExNjQyNzM.*_ga_W0YLR4190T*MTcwMTgzNzYzMy42LjEuMTcwMTgzODU2Mi4wLjAuMA..)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "# colab 사용시 파일 다운로드\n",
    "# try:\n",
    "#     from google.colab import files\n",
    "#     files.download('vectors.tsv')\n",
    "#     files.download('metadata.tsv')\n",
    "# except Exception:\n",
    "#     pass"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "provenance": []
  },
  "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": "3.11.4"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": false,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {
    "height": "calc(100% - 180px)",
    "left": "10px",
    "top": "150px",
    "width": "256px"
   },
   "toc_section_display": false,
   "toc_window_display": false
  },
  "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": 1
}