[mtvs]lej_submit_bert.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "347da3bf",
   "metadata": {},
   "source": [
    "#### Module import & Load Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "1a6c54c9",
   "metadata": {},
   "outputs": [],
   "source": [
    "# 메모리내 변수 제거\n",
    "\n",
    "all = [var for var in globals() if var[0] != \"_\"]   # globals() 목록의 첫글자가 _ 로 시작하지 않는 자료의 리스트만 가져와서\n",
    "for var in all:\n",
    "    del globals()[var]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "7aebeef3",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "\n",
    "import wordcloud\n",
    "import nltk\n",
    "# import seaborn as sns\n",
    "import re\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib.pyplot import figure\n",
    "\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.metrics import f1_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "dc13fc3c-c1b1-4b77-9ea2-8742c3633345",
   "metadata": {},
   "outputs": [],
   "source": [
    "train1 = pd.read_csv('c:/data/project/train.csv')\n",
    "test = pd.read_csv('c:/data/project/test.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "afc501c2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 2478 entries, 0 to 2477\n",
      "Data columns (total 5 columns):\n",
      " #   Column              Non-Null Count  Dtype \n",
      "---  ------              --------------  ----- \n",
      " 0   ID                  2478 non-null   object\n",
      " 1   first_party         2478 non-null   object\n",
      " 2   second_party        2478 non-null   object\n",
      " 3   facts               2478 non-null   object\n",
      " 4   first_party_winner  2478 non-null   int64 \n",
      "dtypes: int64(1), object(4)\n",
      "memory usage: 96.9+ KB\n"
     ]
    },
    {
     "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>ID</th>\n",
       "      <th>first_party</th>\n",
       "      <th>second_party</th>\n",
       "      <th>facts</th>\n",
       "      <th>first_party_winner</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>TRAIN_0000</td>\n",
       "      <td>Phil A. St. Amant</td>\n",
       "      <td>Herman A. Thompson</td>\n",
       "      <td>On June 27, 1962, Phil St. Amant, a candidate ...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>TRAIN_0001</td>\n",
       "      <td>Stephen Duncan</td>\n",
       "      <td>Lawrence Owens</td>\n",
       "      <td>Ramon Nelson was riding his bike when he suffe...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>TRAIN_0002</td>\n",
       "      <td>Billy Joe Magwood</td>\n",
       "      <td>Tony Patterson, Warden, et al.</td>\n",
       "      <td>An Alabama state court convicted Billy Joe Mag...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>TRAIN_0003</td>\n",
       "      <td>Linkletter</td>\n",
       "      <td>Walker</td>\n",
       "      <td>Victor Linkletter was convicted in state court...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>TRAIN_0004</td>\n",
       "      <td>William Earl Fikes</td>\n",
       "      <td>Alabama</td>\n",
       "      <td>On April 24, 1953 in Selma, Alabama, an intrud...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           ID         first_party                    second_party  \\\n",
       "0  TRAIN_0000   Phil A. St. Amant              Herman A. Thompson   \n",
       "1  TRAIN_0001      Stephen Duncan                  Lawrence Owens   \n",
       "2  TRAIN_0002   Billy Joe Magwood  Tony Patterson, Warden, et al.   \n",
       "3  TRAIN_0003          Linkletter                          Walker   \n",
       "4  TRAIN_0004  William Earl Fikes                         Alabama   \n",
       "\n",
       "                                               facts  first_party_winner  \n",
       "0  On June 27, 1962, Phil St. Amant, a candidate ...                   1  \n",
       "1  Ramon Nelson was riding his bike when he suffe...                   0  \n",
       "2  An Alabama state court convicted Billy Joe Mag...                   1  \n",
       "3  Victor Linkletter was convicted in state court...                   0  \n",
       "4  On April 24, 1953 in Selma, Alabama, an intrud...                   1  "
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_nlp = train1.copy()\n",
    "train_nlp.info()\n",
    "train_nlp.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "c689f2b0",
   "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>facts</th>\n",
       "      <th>first_party_winner</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Phil A. St. Amant, Herman A. Thompson, On June...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Stephen Duncan, Lawrence Owens, Ramon Nelson w...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Billy Joe Magwood, Tony Patterson, Warden, et ...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Linkletter, Walker, Victor Linkletter was conv...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>William Earl Fikes, Alabama, On April 24, 1953...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2473</th>\n",
       "      <td>HollyFrontier Cheyenne Refining, LLC, et al., ...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2474</th>\n",
       "      <td>Grupo Mexicano de Desarrollo, S. A., Alliance ...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2475</th>\n",
       "      <td>Peguero, United States, In 1992, the District ...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2476</th>\n",
       "      <td>Immigration and Naturalization Service, St. Cy...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2477</th>\n",
       "      <td>Markman, Westview Instruments, Inc., Herbert M...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2478 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  facts  first_party_winner\n",
       "0     Phil A. St. Amant, Herman A. Thompson, On June...                   1\n",
       "1     Stephen Duncan, Lawrence Owens, Ramon Nelson w...                   0\n",
       "2     Billy Joe Magwood, Tony Patterson, Warden, et ...                   1\n",
       "3     Linkletter, Walker, Victor Linkletter was conv...                   0\n",
       "4     William Earl Fikes, Alabama, On April 24, 1953...                   1\n",
       "...                                                 ...                 ...\n",
       "2473  HollyFrontier Cheyenne Refining, LLC, et al., ...                   1\n",
       "2474  Grupo Mexicano de Desarrollo, S. A., Alliance ...                   1\n",
       "2475  Peguero, United States, In 1992, the District ...                   0\n",
       "2476  Immigration and Naturalization Service, St. Cy...                   0\n",
       "2477  Markman, Westview Instruments, Inc., Herbert M...                   0\n",
       "\n",
       "[2478 rows x 2 columns]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_nlp['facts'] = train_nlp['first_party']+', '+train_nlp['second_party']+', '+train_nlp['facts']\n",
    "train_nlp = train_nlp[['facts','first_party_winner']]\n",
    "train_nlp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "25cdc9ce",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_target = pd.DataFrame(train_nlp, columns=['first_party_winner'])\n",
    "# train_target"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "c35b909e",
   "metadata": {},
   "source": [
    "### 불용어 사전 정의 (사용자 지정 & 기본 제공[english])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "35b83b0b",
   "metadata": {},
   "outputs": [],
   "source": [
    "##### 불용어 처리 및 어간과 표제어 추출 함수 #####\n",
    "\n",
    "def utils_preprocess_text(text, flg_stemm=False, flg_lemm=True, lst_stopwords=None): \n",
    "## clean (convert to lowercase and remove punctuations and characters and then strip)\n",
    "\n",
    "    # text = re.sub(r'[^\\w\\s]', '', str(text).lower().strip())\n",
    "    text = re.sub(r'[^\\w\\s]', '', str(text).strip())\n",
    "\n",
    "    # 정규식, 치환 문자열, src 문자열 입력 -> 해당 문자가 발견되지 않는 경우 space bar로 변환, 단어별 split을 위한 작업\n",
    "            \n",
    "    ## Tokenize (convert from string to list)\n",
    "    lst_text = text.split()    ## remove Stopwords\n",
    "    if lst_stopwords is not None:\n",
    "        lst_text = [word for word in lst_text if word not in \n",
    "                    lst_stopwords]\n",
    "                        \n",
    "    ## Stemming (remove -ing, -ly, ...)\n",
    "    if flg_stemm == True:\n",
    "        ps = nltk.stem.porter.PorterStemmer()\n",
    "        lst_text = [ps.stem(word) for word in lst_text]\n",
    "                \n",
    "    ## Lemmatisation (convert the word into root word)\n",
    "    if flg_lemm == True:\n",
    "        lem = nltk.stem.wordnet.WordNetLemmatizer()\n",
    "        lst_text = [lem.lemmatize(word) for word in lst_text]\n",
    "            \n",
    "    ## back to string from list\n",
    "    text = \" \".join(lst_text)\n",
    "    return text"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "d2095585",
   "metadata": {},
   "outputs": [],
   "source": [
    "##### 단어와 빈도수 측정 및 시각화 함수 #####\n",
    "\n",
    "def word_freq(data, columns, titleName=None, graph=False):\n",
    "    corpus = data[columns]\n",
    "    lst_tokens = nltk.tokenize.word_tokenize(corpus.str.cat(sep=\" \"))\n",
    "\n",
    "    dic_word_freq = nltk.FreqDist(lst_tokens)\n",
    "    words_freq = pd.DataFrame(dic_word_freq.most_common(), columns=[\"Word\",\"Freq\"])\n",
    "\n",
    "    if graph==True:\n",
    "        fig, ax = plt.subplots(nrows=1, ncols=1)\n",
    "        fig.suptitle('Most frequent words'+titleName, fontsize=15)\n",
    "        figure(figsize=(30, 24))\n",
    "        \n",
    "        words_freq.set_index(\"Word\").iloc[:30,:].sort_values(by=\"Freq\").plot(\n",
    "                  kind=\"barh\", title=\"words\", ax=ax, \n",
    "                  legend=False).grid(axis='x')\n",
    "               \n",
    "        ax.set(ylabel=None)\n",
    "        plt.show()\n",
    "    \n",
    "    return [words_freq, dic_word_freq]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "ddec3ea0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 3000x2400 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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>Word</th>\n",
       "      <th>Freq</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>the</td>\n",
       "      <td>32628</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>,</td>\n",
       "      <td>27030</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>.</td>\n",
       "      <td>18818</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>of</td>\n",
       "      <td>15867</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>to</td>\n",
       "      <td>11805</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22340</th>\n",
       "      <td>patentee</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22341</th>\n",
       "      <td>localize</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22342</th>\n",
       "      <td>spurious</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22343</th>\n",
       "      <td>bar-coded</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22344</th>\n",
       "      <td>province</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>22345 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            Word   Freq\n",
       "0            the  32628\n",
       "1              ,  27030\n",
       "2              .  18818\n",
       "3             of  15867\n",
       "4             to  11805\n",
       "...          ...    ...\n",
       "22340   patentee      1\n",
       "22341   localize      1\n",
       "22342   spurious      1\n",
       "22343  bar-coded      1\n",
       "22344   province      1\n",
       "\n",
       "[22345 rows x 2 columns]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "word_freq(train_nlp, 'facts', ':basic', True)[0]"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "1107194d",
   "metadata": {},
   "source": [
    "##### 기본 불용어 처리"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "930260b7",
   "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>facts</th>\n",
       "      <th>first_party_winner</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Phil A St Amant Herman A Thompson On June 27 1...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Stephen Duncan Lawrence Owens Ramon Nelson rid...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Billy Joe Magwood Tony Patterson Warden et al ...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Linkletter Walker Victor Linkletter convicted ...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>William Earl Fikes Alabama On April 24 1953 Se...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2473</th>\n",
       "      <td>HollyFrontier Cheyenne Refining LLC et al Rene...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2474</th>\n",
       "      <td>Grupo Mexicano de Desarrollo S A Alliance Bond...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2475</th>\n",
       "      <td>Peguero United States In 1992 District Court s...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2476</th>\n",
       "      <td>Immigration Naturalization Service St Cyr On M...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2477</th>\n",
       "      <td>Markman Westview Instruments Inc Herbert Markm...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2478 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  facts  first_party_winner\n",
       "0     Phil A St Amant Herman A Thompson On June 27 1...                   1\n",
       "1     Stephen Duncan Lawrence Owens Ramon Nelson rid...                   0\n",
       "2     Billy Joe Magwood Tony Patterson Warden et al ...                   1\n",
       "3     Linkletter Walker Victor Linkletter convicted ...                   0\n",
       "4     William Earl Fikes Alabama On April 24 1953 Se...                   1\n",
       "...                                                 ...                 ...\n",
       "2473  HollyFrontier Cheyenne Refining LLC et al Rene...                   1\n",
       "2474  Grupo Mexicano de Desarrollo S A Alliance Bond...                   1\n",
       "2475  Peguero United States In 1992 District Court s...                   0\n",
       "2476  Immigration Naturalization Service St Cyr On M...                   0\n",
       "2477  Markman Westview Instruments Inc Herbert Markm...                   0\n",
       "\n",
       "[2478 rows x 2 columns]"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lst_stopwords = nltk.corpus.stopwords.words(\"english\")\n",
    "# NLTK가 정의한 영어 불용어 리스트를 리턴\n",
    "\n",
    "train_nlp[\"facts\"] = train_nlp[\"facts\"].apply(lambda x: utils_preprocess_text(x, flg_stemm=False, flg_lemm=True, lst_stopwords=lst_stopwords))\n",
    "# nltk 기본 불용어를 제거한 결과\n",
    "train_nlp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "fc91b3a8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 3000x2400 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "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>Word</th>\n",
       "      <th>Freq</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>The</td>\n",
       "      <td>6302</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>court</td>\n",
       "      <td>4497</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Court</td>\n",
       "      <td>4109</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>district</td>\n",
       "      <td>1994</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Appeals</td>\n",
       "      <td>1869</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21656</th>\n",
       "      <td>patentee</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21657</th>\n",
       "      <td>localize</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21658</th>\n",
       "      <td>spurious</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21659</th>\n",
       "      <td>barcoded</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21660</th>\n",
       "      <td>Westviews</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>21661 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "            Word  Freq\n",
       "0            The  6302\n",
       "1          court  4497\n",
       "2          Court  4109\n",
       "3       district  1994\n",
       "4        Appeals  1869\n",
       "...          ...   ...\n",
       "21656   patentee     1\n",
       "21657   localize     1\n",
       "21658   spurious     1\n",
       "21659   barcoded     1\n",
       "21660  Westviews     1\n",
       "\n",
       "[21661 rows x 2 columns]"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "word_freq(train_nlp, 'facts', ':remove basic stop_words',True)[0]"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "27395ca4",
   "metadata": {},
   "source": [
    "#### 승소, 패소 시의 단어 빈도수 확인"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "28f9db67",
   "metadata": {},
   "source": [
    "###### 위의 전처리 과정에서 얻은 facts와 first_party_winner만 남긴 데이터 프레임을 train_nlp2로 지정함"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "1bdcab06",
   "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>facts</th>\n",
       "      <th>first_party_winner</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Phil A St Amant Herman A Thompson On June 27 1...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Stephen Duncan Lawrence Owens Ramon Nelson rid...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Billy Joe Magwood Tony Patterson Warden et al ...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Linkletter Walker Victor Linkletter convicted ...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>William Earl Fikes Alabama On April 24 1953 Se...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2473</th>\n",
       "      <td>HollyFrontier Cheyenne Refining LLC et al Rene...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2474</th>\n",
       "      <td>Grupo Mexicano de Desarrollo S A Alliance Bond...</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2475</th>\n",
       "      <td>Peguero United States In 1992 District Court s...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2476</th>\n",
       "      <td>Immigration Naturalization Service St Cyr On M...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2477</th>\n",
       "      <td>Markman Westview Instruments Inc Herbert Markm...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2478 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  facts  first_party_winner\n",
       "0     Phil A St Amant Herman A Thompson On June 27 1...                   1\n",
       "1     Stephen Duncan Lawrence Owens Ramon Nelson rid...                   0\n",
       "2     Billy Joe Magwood Tony Patterson Warden et al ...                   1\n",
       "3     Linkletter Walker Victor Linkletter convicted ...                   0\n",
       "4     William Earl Fikes Alabama On April 24 1953 Se...                   1\n",
       "...                                                 ...                 ...\n",
       "2473  HollyFrontier Cheyenne Refining LLC et al Rene...                   1\n",
       "2474  Grupo Mexicano de Desarrollo S A Alliance Bond...                   1\n",
       "2475  Peguero United States In 1992 District Court s...                   0\n",
       "2476  Immigration Naturalization Service St Cyr On M...                   0\n",
       "2477  Markman Westview Instruments Inc Herbert Markm...                   0\n",
       "\n",
       "[2478 rows x 2 columns]"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# train_nlp = pd.concat([train_nlp,train_target['first_party_winner']],axis=1, join='inner')\n",
    "train_nlp"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "7e8b3119-8c8b-47e9-a73b-80513b6d07f0",
   "metadata": {},
   "source": [
    "## Define Model & Train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "44f90487",
   "metadata": {},
   "outputs": [],
   "source": [
    "vectorizer = TfidfVectorizer()\n",
    "X = vectorizer.fit_transform(train_nlp['facts'])\n",
    "y = train_nlp['first_party_winner']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "id": "2ad703be",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2478, 19794)"
      ]
     },
     "execution_count": 139,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "id": "9b8854b3",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_eval, y_train, y_eval = train_test_split(X, y, test_size=0.3, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "id": "d7876532",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-6 {color: black;background-color: white;}#sk-container-id-6 pre{padding: 0;}#sk-container-id-6 div.sk-toggleable {background-color: white;}#sk-container-id-6 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-6 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-6 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-6 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-6 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-6 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-6 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-6 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-6 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-6 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-6 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-6 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-6 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-6 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-6 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-6 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-6 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-6 div.sk-item {position: relative;z-index: 1;}#sk-container-id-6 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-6 div.sk-item::before, #sk-container-id-6 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-6 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-6 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-6 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-6 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-6 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-6 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-6 div.sk-label-container {text-align: center;}#sk-container-id-6 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-6 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-6\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>LogisticRegression()</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-6\" type=\"checkbox\" checked><label for=\"sk-estimator-id-6\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">LogisticRegression</label><div class=\"sk-toggleable__content\"><pre>LogisticRegression()</pre></div></div></div></div></div>"
      ],
      "text/plain": [
       "LogisticRegression()"
      ]
     },
     "execution_count": 141,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lr = LogisticRegression()\n",
    "lr.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "id": "5951e5b5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LR Accuracy : 0.674731\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.8038897893030793"
      ]
     },
     "execution_count": 142,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(\"LR Accuracy : %f\" % lr.score(X_eval, y_eval)) # 검증 데이터로 성능 측정\n",
    "accuracy_score(y_eval, lr.predict(X_eval))\n",
    "\n",
    "f1_score(y_eval, lr.predict(X_eval))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "id": "7ff764ed",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1    735\n",
       "0      9\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 143,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_lr = pd.DataFrame(lr.predict(X_eval))\n",
    "test_lr.value_counts()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "699a5a20",
   "metadata": {},
   "source": [
    "### Test csv 적용"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "id": "72d3cb31",
   "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>ID</th>\n",
       "      <th>first_party</th>\n",
       "      <th>second_party</th>\n",
       "      <th>facts</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>TEST_0000</td>\n",
       "      <td>Salerno</td>\n",
       "      <td>United States</td>\n",
       "      <td>The 1984 Bail Reform Act allowed the federal c...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>TEST_0001</td>\n",
       "      <td>Milberg Weiss Bershad Hynes and Lerach</td>\n",
       "      <td>Lexecon, Inc.</td>\n",
       "      <td>Lexecon Inc. was a defendant in a class action...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>TEST_0002</td>\n",
       "      <td>No. 07-582\\t Title: \\t Federal Communications ...</td>\n",
       "      <td>Fox Television Stations, Inc., et al.</td>\n",
       "      <td>In 2002 and 2003, Fox Television Stations broa...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>TEST_0003</td>\n",
       "      <td>Harold Kaufman</td>\n",
       "      <td>United States</td>\n",
       "      <td>During his trial for armed robbery of a federa...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>TEST_0004</td>\n",
       "      <td>Berger</td>\n",
       "      <td>Hanlon</td>\n",
       "      <td>In 1993, a magistrate judge issued a warrant a...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          ID                                        first_party  \\\n",
       "0  TEST_0000                                            Salerno   \n",
       "1  TEST_0001             Milberg Weiss Bershad Hynes and Lerach   \n",
       "2  TEST_0002  No. 07-582\\t Title: \\t Federal Communications ...   \n",
       "3  TEST_0003                                    Harold Kaufman    \n",
       "4  TEST_0004                                             Berger   \n",
       "\n",
       "                            second_party  \\\n",
       "0                          United States   \n",
       "1                          Lexecon, Inc.   \n",
       "2  Fox Television Stations, Inc., et al.   \n",
       "3                          United States   \n",
       "4                                 Hanlon   \n",
       "\n",
       "                                               facts  \n",
       "0  The 1984 Bail Reform Act allowed the federal c...  \n",
       "1  Lexecon Inc. was a defendant in a class action...  \n",
       "2  In 2002 and 2003, Fox Television Stations broa...  \n",
       "3  During his trial for armed robbery of a federa...  \n",
       "4  In 1993, a magistrate judge issued a warrant a...  "
      ]
     },
     "execution_count": 144,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "id": "8b156950",
   "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>ID</th>\n",
       "      <th>first_party</th>\n",
       "      <th>second_party</th>\n",
       "      <th>facts</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>TEST_0000</td>\n",
       "      <td>Salerno</td>\n",
       "      <td>United States</td>\n",
       "      <td>Salerno, United States, The 1984 Bail Reform A...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>TEST_0001</td>\n",
       "      <td>Milberg Weiss Bershad Hynes and Lerach</td>\n",
       "      <td>Lexecon, Inc.</td>\n",
       "      <td>Milberg Weiss Bershad Hynes and Lerach, Lexeco...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>TEST_0002</td>\n",
       "      <td>No. 07-582\\t Title: \\t Federal Communications ...</td>\n",
       "      <td>Fox Television Stations, Inc., et al.</td>\n",
       "      <td>No. 07-582\\t Title: \\t Federal Communications ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>TEST_0003</td>\n",
       "      <td>Harold Kaufman</td>\n",
       "      <td>United States</td>\n",
       "      <td>Harold Kaufman , United States, During his tri...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>TEST_0004</td>\n",
       "      <td>Berger</td>\n",
       "      <td>Hanlon</td>\n",
       "      <td>Berger, Hanlon, In 1993, a magistrate judge is...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1235</th>\n",
       "      <td>TEST_1235</td>\n",
       "      <td>Haitian Centers Council, Inc., et al.</td>\n",
       "      <td>Chris Sale, Acting Commissioner, Immigration A...</td>\n",
       "      <td>Haitian Centers Council, Inc., et al., Chris S...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1236</th>\n",
       "      <td>TEST_1236</td>\n",
       "      <td>Whitman</td>\n",
       "      <td>American Trucking Associations, Inc.</td>\n",
       "      <td>Whitman, American Trucking Associations, Inc.,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1237</th>\n",
       "      <td>TEST_1237</td>\n",
       "      <td>Linda A. Matteo and John J. Madigan</td>\n",
       "      <td>William G. Barr</td>\n",
       "      <td>Linda A. Matteo and John J. Madigan, William G...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1238</th>\n",
       "      <td>TEST_1238</td>\n",
       "      <td>Washington State Apple Advertising Commission</td>\n",
       "      <td>Hunt</td>\n",
       "      <td>Washington State Apple Advertising Commission,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1239</th>\n",
       "      <td>TEST_1239</td>\n",
       "      <td>Theodore Stovall</td>\n",
       "      <td>Wilfred Denno, Warden</td>\n",
       "      <td>Theodore Stovall, Wilfred Denno, Warden, On Au...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1240 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "             ID                                        first_party  \\\n",
       "0     TEST_0000                                            Salerno   \n",
       "1     TEST_0001             Milberg Weiss Bershad Hynes and Lerach   \n",
       "2     TEST_0002  No. 07-582\\t Title: \\t Federal Communications ...   \n",
       "3     TEST_0003                                    Harold Kaufman    \n",
       "4     TEST_0004                                             Berger   \n",
       "...         ...                                                ...   \n",
       "1235  TEST_1235              Haitian Centers Council, Inc., et al.   \n",
       "1236  TEST_1236                                            Whitman   \n",
       "1237  TEST_1237                Linda A. Matteo and John J. Madigan   \n",
       "1238  TEST_1238      Washington State Apple Advertising Commission   \n",
       "1239  TEST_1239                                   Theodore Stovall   \n",
       "\n",
       "                                           second_party  \\\n",
       "0                                         United States   \n",
       "1                                         Lexecon, Inc.   \n",
       "2                 Fox Television Stations, Inc., et al.   \n",
       "3                                         United States   \n",
       "4                                                Hanlon   \n",
       "...                                                 ...   \n",
       "1235  Chris Sale, Acting Commissioner, Immigration A...   \n",
       "1236               American Trucking Associations, Inc.   \n",
       "1237                                    William G. Barr   \n",
       "1238                                               Hunt   \n",
       "1239                              Wilfred Denno, Warden   \n",
       "\n",
       "                                                  facts  \n",
       "0     Salerno, United States, The 1984 Bail Reform A...  \n",
       "1     Milberg Weiss Bershad Hynes and Lerach, Lexeco...  \n",
       "2     No. 07-582\\t Title: \\t Federal Communications ...  \n",
       "3     Harold Kaufman , United States, During his tri...  \n",
       "4     Berger, Hanlon, In 1993, a magistrate judge is...  \n",
       "...                                                 ...  \n",
       "1235  Haitian Centers Council, Inc., et al., Chris S...  \n",
       "1236  Whitman, American Trucking Associations, Inc.,...  \n",
       "1237  Linda A. Matteo and John J. Madigan, William G...  \n",
       "1238  Washington State Apple Advertising Commission,...  \n",
       "1239  Theodore Stovall, Wilfred Denno, Warden, On Au...  \n",
       "\n",
       "[1240 rows x 4 columns]"
      ]
     },
     "execution_count": 145,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test['facts'] = test['first_party']+', '+test['second_party']+', '+test['facts']\n",
    "# test = test['facts']\n",
    "test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "id": "9dd2f5a1",
   "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>facts</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Salerno, United States, The 1984 Bail Reform A...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Milberg Weiss Bershad Hynes and Lerach, Lexeco...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>No. 07-582\\t Title: \\t Federal Communications ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Harold Kaufman , United States, During his tri...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Berger, Hanlon, In 1993, a magistrate judge is...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1235</th>\n",
       "      <td>Haitian Centers Council, Inc., et al., Chris S...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1236</th>\n",
       "      <td>Whitman, American Trucking Associations, Inc.,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1237</th>\n",
       "      <td>Linda A. Matteo and John J. Madigan, William G...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1238</th>\n",
       "      <td>Washington State Apple Advertising Commission,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1239</th>\n",
       "      <td>Theodore Stovall, Wilfred Denno, Warden, On Au...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1240 rows × 1 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  facts\n",
       "0     Salerno, United States, The 1984 Bail Reform A...\n",
       "1     Milberg Weiss Bershad Hynes and Lerach, Lexeco...\n",
       "2     No. 07-582\\t Title: \\t Federal Communications ...\n",
       "3     Harold Kaufman , United States, During his tri...\n",
       "4     Berger, Hanlon, In 1993, a magistrate judge is...\n",
       "...                                                 ...\n",
       "1235  Haitian Centers Council, Inc., et al., Chris S...\n",
       "1236  Whitman, American Trucking Associations, Inc.,...\n",
       "1237  Linda A. Matteo and John J. Madigan, William G...\n",
       "1238  Washington State Apple Advertising Commission,...\n",
       "1239  Theodore Stovall, Wilfred Denno, Warden, On Au...\n",
       "\n",
       "[1240 rows x 1 columns]"
      ]
     },
     "execution_count": 146,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test = pd.DataFrame(test, columns=['facts'])\n",
    "test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "id": "1aa68c88",
   "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>facts</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Salerno United States The 1984 Bail Reform Act...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Milberg Weiss Bershad Hynes and Lerach Lexecon...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>No 07582 Title Federal Communications Commissi...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Harold Kaufman United States During his trial ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Berger Hanlon In 1993 a magistrate judge issue...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1235</th>\n",
       "      <td>Haitian Centers Council Inc et al Chris Sale A...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1236</th>\n",
       "      <td>Whitman American Trucking Associations Inc Sec...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1237</th>\n",
       "      <td>Linda A Matteo and John J Madigan William G Ba...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1238</th>\n",
       "      <td>Washington State Apple Advertising Commission ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1239</th>\n",
       "      <td>Theodore Stovall Wilfred Denno Warden On Augus...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1240 rows × 1 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  facts\n",
       "0     Salerno United States The 1984 Bail Reform Act...\n",
       "1     Milberg Weiss Bershad Hynes and Lerach Lexecon...\n",
       "2     No 07582 Title Federal Communications Commissi...\n",
       "3     Harold Kaufman United States During his trial ...\n",
       "4     Berger Hanlon In 1993 a magistrate judge issue...\n",
       "...                                                 ...\n",
       "1235  Haitian Centers Council Inc et al Chris Sale A...\n",
       "1236  Whitman American Trucking Associations Inc Sec...\n",
       "1237  Linda A Matteo and John J Madigan William G Ba...\n",
       "1238  Washington State Apple Advertising Commission ...\n",
       "1239  Theodore Stovall Wilfred Denno Warden On Augus...\n",
       "\n",
       "[1240 rows x 1 columns]"
      ]
     },
     "execution_count": 147,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test[\"facts\"] = test[\"facts\"].apply(lambda x: utils_preprocess_text(x, flg_stemm=False, flg_lemm=True))\n",
    "test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "id": "11375ee6",
   "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>facts</th>\n",
       "      <th>first_party_winner</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Salerno United States The 1984 Bail Reform Act...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Milberg Weiss Bershad Hynes Lerach Lexecon Inc...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>No 07582 Title Federal Communications Commissi...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Harold Kaufman United States During trial arme...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Berger Hanlon In 1993 magistrate judge issued ...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1235</th>\n",
       "      <td>Haitian Centers Council Inc et al Chris Sale A...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1236</th>\n",
       "      <td>Whitman American Trucking Associations Inc Sec...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1237</th>\n",
       "      <td>Linda A Matteo John J Madigan William G Barr L...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1238</th>\n",
       "      <td>Washington State Apple Advertising Commission ...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1239</th>\n",
       "      <td>Theodore Stovall Wilfred Denno Warden On Augus...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1240 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                  facts  first_party_winner\n",
       "0     Salerno United States The 1984 Bail Reform Act...                   0\n",
       "1     Milberg Weiss Bershad Hynes Lerach Lexecon Inc...                   0\n",
       "2     No 07582 Title Federal Communications Commissi...                   0\n",
       "3     Harold Kaufman United States During trial arme...                   0\n",
       "4     Berger Hanlon In 1993 magistrate judge issued ...                   0\n",
       "...                                                 ...                 ...\n",
       "1235  Haitian Centers Council Inc et al Chris Sale A...                   0\n",
       "1236  Whitman American Trucking Associations Inc Sec...                   0\n",
       "1237  Linda A Matteo John J Madigan William G Barr L...                   0\n",
       "1238  Washington State Apple Advertising Commission ...                   0\n",
       "1239  Theodore Stovall Wilfred Denno Warden On Augus...                   0\n",
       "\n",
       "[1240 rows x 2 columns]"
      ]
     },
     "execution_count": 148,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dfTest = pd.DataFrame(test['facts'])\n",
    "dfTest[\"facts\"] = dfTest[\"facts\"].apply(lambda x: utils_preprocess_text(x, flg_stemm=False, flg_lemm=True, lst_stopwords=lst_stopwords))\n",
    "dfTest['first_party_winner'] = np.zeros(len(dfTest)).astype(int)\n",
    "dfTest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "id": "0c3329bf",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<1240x19794 sparse matrix of type '<class 'numpy.float64'>'\n",
       "\twith 91260 stored elements in Compressed Sparse Row format>"
      ]
     },
     "execution_count": 149,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test = vectorizer.transform(dfTest['facts'])\n",
    "y_test = dfTest['first_party_winner']\n",
    "X_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "id": "a43b68e3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "first_party_winner\n",
       "1                     1214\n",
       "0                       26\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 151,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred_test = lr.predict(X_test)\n",
    "predcsv = pd.DataFrame(y_pred_test,columns=['first_party_winner'])\n",
    "predcsv.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b3595fd2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "first_party_winner\n",
       "1                     1240\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred_test = automl.predict((np.asarray(X_test)))\n",
    "predcsv = pd.DataFrame(y_pred_test,columns=['first_party_winner'])\n",
    "predcsv.value_counts()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "id": "a4aa8041-458a-4404-8bb1-d0169d4dd64d",
   "metadata": {},
   "source": [
    "## Inference & Submission"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "id": "7b3f57bb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Done\n"
     ]
    }
   ],
   "source": [
    "submit = pd.read_csv('C:/data/project/sample_submission.csv')\n",
    "submit['first_party_winner'] = predcsv\n",
    "submit.to_csv('./sample_submission.csv', index=False)\n",
    "print('Done')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "PyTorch 1.14 (NGC 22.12/Python 3.8) on Backend.AI",
   "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": "3.11.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}