notebooks/second_model.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 두 번째 모델 훈련하기\n",
    "\n",
    "이 노트북에서 첫 번째 모델의 단점을 극복하기 위한 특성을 사용해 두 번째 모델을 훈련합니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package vader_lexicon to\n",
      "[nltk_data]     /home/haesun/nltk_data...\n",
      "[nltk_data]   Package vader_lexicon is already up-to-date!\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from pathlib import Path\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.metrics import accuracy_score, f1_score, precision_score, recall_score, classification_report\n",
    "import joblib\n",
    "import sys\n",
    "sys.path.append(\"..\")\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "\n",
    "from ml_editor.data_processing import (\n",
    "    format_raw_df,\n",
    "    get_split_by_author,\n",
    "    train_vectorizer, \n",
    "    get_vectorized_series,\n",
    "    get_feature_vector_and_label\n",
    ")\n",
    "from ml_editor.model_evaluation import (\n",
    "    get_feature_importance,\n",
    "    get_roc_plot,\n",
    "    get_confusion_matrix_plot,\n",
    "    get_calibration_plot\n",
    ")\n",
    "\n",
    "from ml_editor.model_v2 import (\n",
    "    add_char_count_features, \n",
    "    get_word_stats, \n",
    "    get_sentiment_score, \n",
    "    POS_NAMES,\n",
    "    get_question_score_from_input\n",
    ")\n",
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "np.random.seed(35)\n",
    "\n",
    "data_path = Path('../data/writers.csv')\n",
    "df = pd.read_csv(data_path)\n",
    "df = format_raw_df(df.copy())\n",
    "df = df.loc[df[\"is_question\"]].copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "df[\"full_text\"] = df[\"Title\"].str.cat(df[\"body_text\"], sep=\" \", na_rep=\"\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "새로운 모델에 도움이 될만한 후보 특성을 추가합니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_df, test_df = get_split_by_author(df, test_size=0.2, random_state=40)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "vectorizer = train_vectorizer(train_df)\n",
    "df[\"vectors\"] = get_vectorized_series(df[\"full_text\"].copy(), vectorizer)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`ml_editor` 소스 코드를 확인하여 이 함수들의 기능을 확인해 보세요!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|████████████████████████████████████████████████████████████████████████████████████████████████████| 7971/7971 [04:03<00:00, 32.70it/s]\n",
      "100%|███████████████████████████████████████████████████████████████████████████████████████████████████| 7971/7971 [00:16<00:00, 495.49it/s]\n"
     ]
    }
   ],
   "source": [
    "df = add_char_count_features(df.copy())\n",
    "df = get_word_stats(df.copy())\n",
    "df = get_sentiment_score(df.copy())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "feature_arr = [\"num_questions\", \n",
    "               \"num_periods\",\n",
    "               \"num_commas\",\n",
    "               \"num_exclam\",\n",
    "               \"num_quotes\",\n",
    "               \"num_colon\",\n",
    "               \"num_stops\",\n",
    "               \"num_semicolon\",\n",
    "               \"num_words\",\n",
    "               \"num_chars\",\n",
    "               \"num_diff_words\",\n",
    "               \"avg_word_len\",\n",
    "               \"polarity\"\n",
    "              ]\n",
    "feature_arr.extend(POS_NAMES.keys())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 모델\n",
    "\n",
    "이제 새로운 특성을 추가했으니 새 모델을 훈련해 보죠. 특성만 다르고 이전과 동일한 모델을 사용합니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 특성을 새로 추가했으므로 데이터셋을 다시 나눕니다.\n",
    "train_df, test_df = get_split_by_author(df, test_size=0.2, random_state=40)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, y_train = get_feature_vector_and_label(train_df, feature_arr)\n",
    "X_test, y_test = get_feature_vector_and_label(test_df, feature_arr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "False    3483\n",
       "True     2959\n",
       "Name: Score, dtype: int64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1529, 7799)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "clf = RandomForestClassifier(n_estimators=100, class_weight='balanced', oob_score=True)\n",
    "clf.fit(X_train, y_train)\n",
    "\n",
    "y_predicted = clf.predict(X_test)\n",
    "y_predicted_proba = clf.predict_proba(X_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, we can measure performance as we saw in the first training notebook."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "훈련 정확도 = 0.594, 정밀도 = 0.571, recall = 0.466, f1 = 0.513\n"
     ]
    }
   ],
   "source": [
    "def get_metrics(y_test, y_predicted):  \n",
    "    # 진짜 양성 / (진짜 양성 + 가짜 양성)\n",
    "    precision = precision_score(y_test, y_predicted, pos_label=True,\n",
    "                                    average='binary')             \n",
    "    # 진짜 양성 / (진짜 양성 + 가짜 음성)\n",
    "    recall = recall_score(y_test, y_predicted, pos_label=True,\n",
    "                              average='binary')\n",
    "    \n",
    "    # 정밀도와 재현율의 조화 평균\n",
    "    f1 = f1_score(y_test, y_predicted, pos_label=True, average='binary')\n",
    "    \n",
    "    # 진짜 양성 + 진짜 음성 / 전체\n",
    "    accuracy = accuracy_score(y_test, y_predicted)\n",
    "    return accuracy, precision, recall, f1\n",
    "\n",
    "\n",
    "# 훈련 정확도\n",
    "# https://datascience.stackexchange.com/questions/13151/randomforestclassifier-oob-scoring-method 참고\n",
    "y_train_pred = np.argmax(clf.oob_decision_function_,axis=1)\n",
    "\n",
    "accuracy, precision, recall, f1 = get_metrics(y_train, y_train_pred)\n",
    "print(\"훈련 정확도 = %.3f, 정밀도 = %.3f, recall = %.3f, f1 = %.3f\" % (accuracy, precision, recall, f1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "검증 정확도 = 0.603, 정밀도 = 0.620, recall = 0.488, f1 = 0.546\n"
     ]
    }
   ],
   "source": [
    "accuracy, precision, recall, f1 = get_metrics(y_test, y_predicted)\n",
    "print(\"검증 정확도 = %.3f, 정밀도 = %.3f, recall = %.3f, f1 = %.3f\" % (accuracy, precision, recall, f1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "다행히 이 모델은 이전 모델보다 성능이 더 높습니다! 새로운 모델과 벡터화 객체를 나중에 사용하기 위해 디스크에 저장하겠습니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['../models/vectorizer_2.pkl']"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_path = Path(\"../models/model_2.pkl\")\n",
    "vectorizer_path = Path(\"../models/vectorizer_2.pkl\")\n",
    "joblib.dump(clf, model_path) \n",
    "joblib.dump(vectorizer, vectorizer_path) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 특성의 유용성 검증하기\n",
    "\n",
    "그다음 특성 중요도 [노트북](https://github.com/rickiepark/ml-powered-applications/blob/master/notebooks/feature_importance.ipynb)에서 설명한 방법을 사용해 새로운 특성을 새로운 모델이 사용하는지 확인해 보겠습니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "w_indices = vectorizer.get_feature_names()\n",
    "w_indices.extend(feature_arr)\n",
    "all_feature_names = np.array(w_indices)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "상위 20개 중요도:\n",
      "\n",
      "num_chars: 0.0085\n",
      "ADJ: 0.0079\n",
      "num_periods: 0.0078\n",
      "num_diff_words: 0.0075\n",
      "num_questions: 0.0073\n",
      "PUNCT: 0.0071\n",
      "num_commas: 0.007\n",
      "ADV: 0.0068\n",
      "AUX: 0.0065\n",
      "num_words: 0.0065\n",
      "DET: 0.0063\n",
      "avg_word_len: 0.0062\n",
      "NOUN: 0.0061\n",
      "polarity: 0.006\n",
      "PRON: 0.0059\n",
      "SCONJ: 0.0058\n",
      "VERB: 0.0057\n",
      "PART: 0.0056\n",
      "num_stops: 0.0056\n",
      "ADP: 0.0054\n",
      "\n",
      "하위 20개 중요도:\n",
      "\n",
      "integration: 0\n",
      "avid: 0\n",
      "obnoxious: 0\n",
      "glare: 0\n",
      "twin: 0\n",
      "twins: 0\n",
      "milk: 0\n",
      "clutter: 0\n",
      "researchers: 0\n",
      "historically: 0\n",
      "coast: 0\n",
      "foil: 0\n",
      "coat: 0\n",
      "authorities: 0\n",
      "inter: 0\n",
      "steel: 0\n",
      "historian: 0\n",
      "possession: 0\n",
      "observed: 0\n",
      "blocked: 0\n"
     ]
    }
   ],
   "source": [
    "k = 20\n",
    "print(\"상위 %s개 중요도:\\n\" % k)\n",
    "print('\\n'.join([\"%s: %.2g\" % (tup[0], tup[1]) for tup in get_feature_importance(clf, all_feature_names)[:k]]))\n",
    "\n",
    "print(\"\\n하위 %s개 중요도:\\n\" % k)\n",
    "print('\\n'.join([\"%s: %.2g\" % (tup[0], tup[1]) for tup in get_feature_importance(clf, all_feature_names)[-k:]]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "새로운 특성이 가장 예측 성능이 좋은 편이군요! 반대로 TF-IDF 벡터화로 얻은 단어 벡터는 특별히 도움이 되는 것 같지 않습니다. 이어지는 [노트북](https://github.com/rickiepark/ml-powered-applications/blob/master/notebooks/third_model.ipynb)에서 이런 특성을 제외하고 세 번째 모델을 훈련하여 어떤 성능을 내는지 확인해 보습니다."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 예측과 데이터 비교하기\n",
    "\n",
    "이 섹션은 새로운 모델로 데이터와 예측 비교하기 [노트북](https://github.com/rickiepark/ml-powered-applications/blob/master/notebooks/comparing_data_to_predictions.ipynb)에서 설명한 평가 방법을 사용합니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "get_roc_plot(y_predicted_proba[:,1], y_test, figsize=(10,10))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x648 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "get_confusion_matrix_plot(y_predicted, y_test, figsize=(9, 9))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "No handles with labels found to put in legend.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\tBrier: 0.235\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x648 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "get_calibration_plot(y_predicted_proba[:,1], y_test, figsize=(9,9))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 추론 함수\n",
    "\n",
    "첫 번째 모델과 마찬가지로 임의의 질문을 받고 높은 점수를 받을 추정 확률을 출력하는 추론 함수를 정의합니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 107.98it/s]\n",
      "100%|████████████████████████████████████████████████████████████████████████████████████████████████████████| 1/1 [00:00<00:00, 1104.93it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.36 probability of the question receiving a high score according to our model\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "pos_prob = get_question_score_from_input(\"\"\"\n",
    "When quoting a person's informal speech, how much liberty do you have to make changes to what they say?\n",
    "\"\"\")\n",
    "\n",
    "print(\"%s probability of the question receiving a high score according to our model\" % (pos_prob))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}