final.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Write a categorizer\n",
    "\n",
    "## Wooheum Xin\n",
    "\n",
    "13 August 2022\n",
    "\n",
    "Introduction to Computational Literary Analysis\n",
    "\n",
    "Department of Digital Humanities, University of California, Berkeley\n",
    "\n",
    "Instructor: Jonathan Reeve\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Users\\xin98\\.conda\\envs\\notebook\\lib\\site-packages\\scipy\\__init__.py:146: UserWarning: A NumPy version >=1.16.5 and <1.23.0 is required for this version of SciPy (detected version 1.23.1\n",
      "  warnings.warn(f\"A NumPy version >={np_minversion} and <{np_maxversion}\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['SimSun']\n"
     ]
    }
   ],
   "source": [
    "import spacy\n",
    "import nltk\n",
    "import pandas as pd\n",
    "import sklearn\n",
    "import numpy as np\n",
    "import string\n",
    "import json\n",
    "import requests\n",
    "import random\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "from tqdm.notebook import tqdm\n",
    "from collections import Counter\n",
    "from nltk.corpus import stopwords\n",
    "from sklearn.base import TransformerMixin\n",
    "from sklearn.feature_extraction._stop_words import ENGLISH_STOP_WORDS\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.svm import LinearSVC\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn import metrics\n",
    "nlp = spacy.load('en_core_web_lg')\n",
    "import matplotlib\n",
    "plt.rc('font', family='SimSun')\n",
    "print(plt.rcParams['font.family'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Go to corpus-db to see which subjects are available"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[['Fiction', 1920],\n",
       " ['Short stories', 1581],\n",
       " ['Science fiction', 1283],\n",
       " ['Adventure stories', 766],\n",
       " ['Historical fiction', 634],\n",
       " ['Poetry', 633],\n",
       " ['Love stories', 619],\n",
       " [', ', 557],\n",
       " ['English wit and humor -- Periodicals', 555],\n",
       " ['Conduct of life -- Juvenile fiction', 554],\n",
       " ['Detective and mystery stories', 537],\n",
       " ['Western stories', 432],\n",
       " ['Man-woman relationships -- Fiction', 389],\n",
       " ['England -- Fiction', 309],\n",
       " ['Friendship -- Juvenile fiction', 307]]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "baseURL = \"http://corpus-db.org\" \n",
    "allSubjects = requests.get(baseURL + \"/api/subjects\").text\n",
    "subjects = json.loads(allSubjects)\n",
    "subjects[:15]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Determined by four subjects\n",
    ">'Detective and mystery stories','Adventure stories','Love stories','Science fiction'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "sub_list=['Detective and mystery stories','Adventure stories','Love stories','Science fiction']\n",
    "AllMeta={} \n",
    "def getMeta(sub_list):\n",
    "    for sub in sub_list:\n",
    "        AllMeta[sub]=json.loads(requests.get(baseURL + \"/api/subject/\"+sub).text)\n",
    "getMeta(sub_list)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Appending a fixed number of novels in the Fulltext[]."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def Meta2Fulltext(Meta,num):\n",
    "    Fulltext,usefull_link=[],[]\n",
    "    counter=0\n",
    "    for index in range(len(Meta)):\n",
    "        start_num=Meta[index]['formats'].find('text/plain; charset=utf-8')+29\n",
    "        if(start_num==28):\n",
    "            start_num=Meta[index]['formats'].find(\"text/plain':\")+14\n",
    "            if(start_num==13):\n",
    "                continue\n",
    "        new_str=Meta[index]['formats'][start_num:]\n",
    "        weblink=new_str[:new_str.find(\"'\")]\n",
    "        if(weblink[-3:]=='zip'):\n",
    "            continue\n",
    "        usefull_link.append(weblink)\n",
    "        counter=counter+1\n",
    "        if(counter==num):\n",
    "            break\n",
    "    for index in tqdm(range(len(usefull_link))):\n",
    "        tmptext=requests.get(usefull_link[index])\n",
    "        Fulltext.append(tmptext.text)\n",
    "    return Fulltext"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading :Detective and mystery stories\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "382f4f820a374b6ba2b903976a12edbf",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading :Adventure stories\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "179a49c8655f4179b667d78035b2e150",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading :Love stories\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "876a0254816e4191a02bdaf037f93f49",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading :Science fiction\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "69523f23ecc143728e5a5376ff3f68f1",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "AllText={}\n",
    "for key, value in AllMeta.items():\n",
    "    print(\"Loading :\"+key)\n",
    "    AllText[key]=Meta2Fulltext(value,100)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Clearing any unwanted parts of the text."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "STOPLIST = set(stopwords.words('english') + list(ENGLISH_STOP_WORDS))\n",
    "SYMBOLS = \" \".join(string.punctuation).split(\" \") + [\"-\", \"...\", \"”\", \"”\"]\n",
    "def Get_Text(ALLTexts):\n",
    "    df=pd.DataFrame(columns=['subject','novel'])\n",
    "    for cate,novel in ALLTexts.items():\n",
    "        for text in novel:\n",
    "            end_num=text.find(\"*** END OF \")\n",
    "            start_num=text.find(\"*** START OF \")+3\n",
    "            tmp=text[start_num:].find(\"***\")+3\n",
    "            finalText=text[start_num+tmp:end_num]\n",
    "            finalText = finalText.strip().replace(\"\\n\", \" \").replace(\"\\r\", \" \")\n",
    "            finalText = finalText.lower()\n",
    "            df.loc[len(df.index)]=[cate,finalText ]\n",
    "    return df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Separating 33% of the dataset for testing purposes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Data Shape: (268, 2)\n",
      "Testing Data Shape: (132, 2)\n"
     ]
    }
   ],
   "source": [
    "AllCleanText=Get_Text(AllText)\n",
    "train, test = train_test_split(AllCleanText, test_size=0.33, random_state=42)\n",
    "print('Training Data Shape:', train.shape)\n",
    "print('Testing Data Shape:', test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Use spaCy for text preprocessing, and then try to find the most used word in each subject"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d58e6cec205f4f2a8b00658438ee98b5",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/61 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7d79a9746d1247038baea4d2dd544b3e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/70 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7c8ccbd38d3149d092ebef5f86e4db67",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/69 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "79ff8764b9d44bacbdef66ff844c8abf",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/68 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "nlp.max_length = 20000000\n",
    "punctuations = string.punctuation\n",
    "def cleanup_text(docs, logging=True):\n",
    "    texts = []\n",
    "    counter = 1\n",
    "    pbar=tqdm(docs)\n",
    "    for doc in pbar:\n",
    "        if counter % 1000 != 0 and logging:\n",
    "            pbar.set_description(\"Processing %s\" %counter)\n",
    "            # print(\"Processed %d out of %d documents.\" % (counter, len(docs)))\n",
    "        counter += 1\n",
    "        doc = nlp(doc, disable=['parser', 'ner'])\n",
    "        tokens = [tok.lemma_.lower().strip() for tok in doc if tok.lemma_ != '-PRON-']\n",
    "        tokens = [tok for tok in tokens if tok not in STOPLIST and tok not in punctuations]\n",
    "        tokens = ' '.join(tokens)\n",
    "        texts.append(tokens)\n",
    "    return pd.Series(texts)\n",
    "clean_text={}\n",
    "for sub in sub_list:\n",
    "    clean_text[sub]=[text for text in train[train['subject'] == sub]['novel']]\n",
    "    clean_text[sub]=' '.join(cleanup_text(clean_text[sub])).split()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Users\\xin98\\.conda\\envs\\notebook\\lib\\site-packages\\IPython\\core\\pylabtools.py:151: UserWarning: Glyph 128 (\\x80) missing from current font.\n",
      "  fig.canvas.print_figure(bytes_io, **kw)\n",
      "c:\\Users\\xin98\\.conda\\envs\\notebook\\lib\\site-packages\\IPython\\core\\pylabtools.py:151: UserWarning: Glyph 157 (\\x9d) missing from current font.\n",
      "  fig.canvas.print_figure(bytes_io, **kw)\n",
      "c:\\Users\\xin98\\.conda\\envs\\notebook\\lib\\site-packages\\IPython\\core\\pylabtools.py:151: UserWarning: Glyph 156 (\\x9c) missing from current font.\n",
      "  fig.canvas.print_figure(bytes_io, **kw)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1296x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Users\\xin98\\.conda\\envs\\notebook\\lib\\site-packages\\IPython\\core\\pylabtools.py:151: UserWarning: Glyph 128 (\\x80) missing from current font.\n",
      "  fig.canvas.print_figure(bytes_io, **kw)\n",
      "c:\\Users\\xin98\\.conda\\envs\\notebook\\lib\\site-packages\\IPython\\core\\pylabtools.py:151: UserWarning: Glyph 157 (\\x9d) missing from current font.\n",
      "  fig.canvas.print_figure(bytes_io, **kw)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1296x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Users\\xin98\\.conda\\envs\\notebook\\lib\\site-packages\\IPython\\core\\pylabtools.py:151: UserWarning: Glyph 128 (\\x80) missing from current font.\n",
      "  fig.canvas.print_figure(bytes_io, **kw)\n",
      "c:\\Users\\xin98\\.conda\\envs\\notebook\\lib\\site-packages\\IPython\\core\\pylabtools.py:151: UserWarning: Glyph 157 (\\x9d) missing from current font.\n",
      "  fig.canvas.print_figure(bytes_io, **kw)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1296x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Users\\xin98\\.conda\\envs\\notebook\\lib\\site-packages\\IPython\\core\\pylabtools.py:151: UserWarning: Glyph 128 (\\x80) missing from current font.\n",
      "  fig.canvas.print_figure(bytes_io, **kw)\n",
      "c:\\Users\\xin98\\.conda\\envs\\notebook\\lib\\site-packages\\IPython\\core\\pylabtools.py:151: UserWarning: Glyph 157 (\\x9d) missing from current font.\n",
      "  fig.canvas.print_figure(bytes_io, **kw)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1296x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for sub in sub_list:\n",
    "    counts=Counter(clean_text[sub])\n",
    "    common_words=[word[0] for word in counts.most_common(20)]\n",
    "    common_counts= [word[1] for word in counts.most_common(20)]\n",
    "    fig = plt.figure(figsize=(18,6))\n",
    "    sns.barplot(x=common_words, y=common_counts)\n",
    "    plt.title('Most Common Words used in the subject for '+sub)\n",
    "    plt.rc('font', family='Malgun Gothic')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Machine Learning with SpaCy\n",
    "### First we load the spacy model Create language processing objects.Then clean up the text with spaCy.Finally, we define a function to show the most significant features, with the highest correlation coefficients"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "nlp.max_length = 30000000\n",
    "def cleanText_spacy(text):\n",
    "    text = text.strip().replace(\"\\n\", \" \").replace(\"\\r\", \" \")\n",
    "    text = text.lower()\n",
    "    return text\n",
    "class CleanTextTransformer(TransformerMixin):   \n",
    "    def transform(self, X, **transform_params):\n",
    "        return [cleanText_spacy(text) for text in X]\n",
    "    def fit(self, X, y=None, **fit_params):\n",
    "        return self\n",
    "    def get_params(self, deep=True):\n",
    "        return {}\n",
    "def tokenizeText(sample):\n",
    "    tokens = nlp(sample)\n",
    "    lemmas = []\n",
    "    for tok in tokens:\n",
    "        lemmas.append(tok.lemma_.lower().strip() if tok.lemma_ != \"-PRON-\" else tok.lower_)\n",
    "    tokens = [tok for tok in lemmas if tok not in [list(STOPLIST)+SYMBOLS]]\n",
    "    return tokens\n",
    "def printNMostInformative(vectorizer, clf, N):\n",
    "    feature_names = vectorizer.get_feature_names_out()\n",
    "    subject_with_fns = sorted(zip(clf.coef_[0], feature_names))\n",
    "    topClass1 = subject_with_fns[:N]\n",
    "    topClass2 = subject_with_fns[:-(N + 1):-1]\n",
    "    print(\"Class 1 best: \")\n",
    "    for feat in topClass1:\n",
    "        print(feat)\n",
    "    print(\"Class 2 best: \")\n",
    "    for feat in topClass2:\n",
    "        print(feat)\n",
    "vectorizer = CountVectorizer(tokenizer=tokenizeText, ngram_range=(1,2))\n",
    "clf = LinearSVC(max_iter=1500000,verbose=True)\n",
    "pipe = Pipeline([('cleanText', CleanTextTransformer()), ('vectorizer', vectorizer), ('clf', clf)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "# data\n",
    "train1 = train['novel'].tolist()\n",
    "labelsTrain1 = train['subject'].tolist()\n",
    "test1 = test['novel'].tolist()\n",
    "labelsTest1 = test['subject'].tolist()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[LibLinear]"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Users\\xin98\\.conda\\envs\\notebook\\lib\\site-packages\\sklearn\\svm\\_base.py:1206: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Pipeline(steps=[('cleanText',\n",
       "                 <__main__.CleanTextTransformer object at 0x000002113D5A4E20>),\n",
       "                ('vectorizer',\n",
       "                 CountVectorizer(ngram_range=(1, 2),\n",
       "                                 tokenizer=<function tokenizeText at 0x0000021140CE5B80>)),\n",
       "                ('clf', LinearSVC(max_iter=1500000, verbose=True))])"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# train\n",
    "pipe.fit(train1, labelsTrain1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy: 0.6590909090909091\n",
      "Top 10 features used to predict: \n",
      "Class 1 best: \n",
      "(-0.002433847469856842, 'detective')\n",
      "(-0.0024188049136491855, ':')\n",
      "(-0.0024100473245777724, 'in')\n",
      "(-0.0019844925803132844, 'holmes')\n",
      "(-0.0016926928753638727, '. ')\n",
      "(-0.0016855530427168576, 'it')\n",
      "(-0.0016536205219722887, 'read')\n",
      "(-0.0015528777223386594, 'you')\n",
      "(-0.0015312585961045255, ': ')\n",
      "(-0.0014776630946859684, 'elephant')\n",
      "Class 2 best: \n",
      "(0.003736269998201579, ' the')\n",
      "(0.002085738187814468, 'on')\n",
      "(0.0018620159926107229, 'cleggett')\n",
      "(0.0017630686206821303, 'boy')\n",
      "(0.0016264766683094017, 'harry')\n",
      "(0.0015937568512690318, 'lion')\n",
      "(0.0015660868016241503, '. .')\n",
      "(0.0015097318678284123, 'rob')\n",
      "(0.0014806946786881587, 'very')\n",
      "(0.0013985720118322542, 'he')\n"
     ]
    }
   ],
   "source": [
    "# test\n",
    "preds = pipe.predict(test1)\n",
    "print(\"accuracy:\", accuracy_score(labelsTest1, preds))\n",
    "print(\"Top 10 features used to predict: \")\n",
    "printNMostInformative(vectorizer, clf, 10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "pipe = Pipeline([('cleanText', CleanTextTransformer()), ('vectorizer', vectorizer)])\n",
    "transform = pipe.fit_transform(train1, labelsTrain1)\n",
    "vocab = vectorizer.get_feature_names_out()\n",
    "for i in range(len(train1)):\n",
    "    s = \"\"\n",
    "    indexIntoVocab = transform.indices[transform.indptr[i]:transform.indptr[i+1]]\n",
    "    numOccurences = transform.data[transform.indptr[i]:transform.indptr[i+1]]\n",
    "    for idx, num in zip(indexIntoVocab, numOccurences):\n",
    "        s += str((vocab[idx], num))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Calculation accuracy, recall, F1 score:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                               precision    recall  f1-score   support\n",
      "\n",
      "Detective and mystery stories       0.54      0.67      0.60        30\n",
      "            Adventure stories       0.77      0.69      0.73        39\n",
      "                 Love stories       0.64      0.58      0.61        31\n",
      "              Science fiction       0.69      0.69      0.69        32\n",
      "\n",
      "                     accuracy                           0.66       132\n",
      "                    macro avg       0.66      0.66      0.66       132\n",
      "                 weighted avg       0.67      0.66      0.66       132\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn import metrics\n",
    "print(metrics.classification_report(labelsTest1, preds, target_names=sub_list))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.9.12 ('notebook')",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.12"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "98610404e4cea0688f837f10e73ecc4df58f27266ebad8bacdb4a3fb12a86a2e"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}