Term Project - language identification.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sys\n",
    "import nltk\n",
    "import string\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.feature_extraction.text import TfidfTransformer\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from imblearn.over_sampling import RandomOverSampler\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "from sklearn.svm import LinearSVC\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "\n",
    "from sklearn.utils.multiclass import unique_labels\n",
    "from sklearn import metrics\n",
    "from sklearn.metrics import confusion_matrix\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "\n",
    "#DSLCC-v1.0\n",
    "#/Users/sudhirsingh/Desktop/Final NLP Term project/dataset/DSLCC-v1.0/train.txt\n",
    "#/Users/sudhirsingh/Desktop/Final NLP Term project/dataset/DSLCC-v1.0/devel.txt\n",
    "#/Users/sudhirsingh/Desktop/Final NLP Term project/dataset/DSLCC-v1.0/test-gold.txt\n",
    "\n",
    "#/Users/sudhirsingh/Desktop/Final NLP Term project/dataset/DSLCC-v1.0/train-eng.txt\n",
    "#/Users/sudhirsingh/Desktop/Final NLP Term project/dataset/DSLCC-v1.0/devel-eng.txt\n",
    "#/Users/sudhirsingh/Desktop/Final NLP Term project/dataset/DSLCC-v1.0/test-eng-gold.txt\n",
    "\n",
    "#DSLCC-v2.0\n",
    "#/Users/sudhirsingh/Desktop/Final NLP Term project/dataset/DSLCC-v2.0/train.txt\n",
    "#/Users/sudhirsingh/Desktop/Final NLP Term project/dataset/DSLCC-v2.0/devel.txt\n",
    "#/Users/sudhirsingh/Desktop/Final NLP Term project/dataset/DSLCC-v2.0/test-ne-gold.txt\n",
    "#/Users/sudhirsingh/Desktop/Final NLP Term project/dataset/DSLCC-v2.0/test-none-gold.txt\n",
    "\n",
    "#DSLCC-v4.0\n",
    "#/Users/sudhirsingh/Desktop/Final NLP Term project/dataset/DSLCC-v2.0/DSL-TRAIN.txt\n",
    "#/Users/sudhirsingh/Desktop/Final NLP Term project/dataset/DSLCC-v2.0/DSL-DEV.txt\n",
    "#/Users/sudhirsingh/Desktop/Final NLP Term project/dataset/DSLCC-v2.0/DSL-TEST-GOLD.txt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Read train and test file."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Enter train file name:/Users/sudhirsingh/Desktop/Final NLP Term project/dataset/DSLCC-v1.0/train.txt\n",
      "Enter dev test file name:/Users/sudhirsingh/Desktop/Final NLP Term project/dataset/DSLCC-v1.0/devel.txt\n",
      "Enter test file name:/Users/sudhirsingh/Desktop/Final NLP Term project/dataset/DSLCC-v1.0/test-gold.txt\n"
     ]
    }
   ],
   "source": [
    "file_train = input(\"Enter train file name:\")\n",
    "file_dev_test = input(\"Enter dev test file name:\")\n",
    "file_test = input(\"Enter test file name:\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Test on train dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "78\n",
      "78\n",
      "82\n"
     ]
    }
   ],
   "source": [
    "print(len(file_train))\n",
    "print(len(file_dev_test))\n",
    "print(len(file_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Open file and preprocess"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def read_file_and_preprocess(file_name):\n",
    "    data = open(file_name).read()\n",
    "    texts, labels = [], []\n",
    "    for idx, line in enumerate(data.split(\"\\n\")):\n",
    "        line_contents = line.split('\\t')\n",
    "        if any(line_contents):\n",
    "            line = line_contents[0]\n",
    "            line = line.replace(\".\", \"\")\n",
    "            line = line.translate(str.maketrans('','','1234567890'))\n",
    "            line = line.translate(str.maketrans('', '', string.punctuation))\n",
    "            line = line.translate(str.maketrans('', '', '#NE#'))          \n",
    "            line = line.replace('\"', '')\n",
    "            line = line.replace('\\n', '')\n",
    "            line = line.strip()\n",
    "            texts.append(line)\n",
    "            labels.append(line_contents[1])\n",
    "    return texts, labels"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Read split train and dev data in X, y train & dev test."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, y_train = read_file_and_preprocess(file_train)\n",
    "X_dev_test, y_dev_test = read_file_and_preprocess(file_dev_test)\n",
    "# test_texts, test_classes = read_file_and_preprocess(file_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def unique_class_label(y_train):\n",
    "    return list(set(y_train))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_unique_class = unique_class_label(y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['sk', 'my', 'pt-PT', 'es-AR', 'pt-BR', 'cz', 'id', 'hr', 'es-ES', 'en-US', 'en-GB', 'sr', 'bs']\n"
     ]
    }
   ],
   "source": [
    "print(train_unique_class)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "novembra je teretni brod s turskom zastavom sa  ilegalnih imigranata bezbjedno dovučen do grčke obale nakon što je uhvaćen na nemirnom moru u blizini jugoistočne obale Grčke\n",
      "bs\n"
     ]
    }
   ],
   "source": [
    "print(X_train[0])\n",
    "print(y_train[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "godišnja Olta Boka osvojila je glavnu nagradu na  acionalnom festivalu pjesme u Albaniji održanom u Tirani tokom tri dana s početkom u petak  decembra\n",
      "bs\n"
     ]
    }
   ],
   "source": [
    "print(X_dev_test[0])\n",
    "print(y_dev_test[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Function to print each line for each language"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def print_line_each_language(X, y):\n",
    "    unique_lang = set()\n",
    "    for line, l_code in zip(X, y):\n",
    "        if l_code not in unique_lang:\n",
    "            unique_lang.add(l_code)\n",
    "            print(line + \"\\t\" + l_code)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### print each line for each language of training set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# print_line_each_language(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### print each line for each language of dev test set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# print_line_each_language(X_dev_test, y_dev_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Feature extraction and modeling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "def feature_extraction_and_modeling(X_train, X_test, feature=\"tf-idf\", ngram=1):\n",
    "#     feature_type = feature\n",
    "    if feature == \"tf-idf\":\n",
    "        # ngram level tf-idf \n",
    "#         min_df=0.01,max_df = 0.95,\n",
    "        word_vectorizer = CountVectorizer(analyzer='word', ngram_range=(1,ngram), max_df = 0.95, max_features=None)\n",
    "#         word_vectorizer = CountVectorizer(max_features=None)\n",
    "    \n",
    "        word_vectorizer_X_train = word_vectorizer.fit_transform(X_train)\n",
    "        tf_idf_transformer = TfidfTransformer()\n",
    "        X_train_feature_matrix = tf_idf_transformer.fit_transform(word_vectorizer_X_train)\n",
    "        \n",
    "        word_vectorizer_X_test = word_vectorizer.transform(X_test)\n",
    "        tf_idf_transformer = TfidfTransformer()\n",
    "        X_test_feature_matrix = tf_idf_transformer.fit_transform(word_vectorizer_X_test)\n",
    "        \n",
    "        \n",
    "#         count_vector = CountVectorizer()\n",
    "#         # create tf-idf vector object \n",
    "#         tf_idf_vector_ngram = TfidfVectorizer(analyzer='word', ngram_range=(1,ngram), max_df=0.95, max_features=None)\n",
    "#         # fit (create matrix) the tf-idf vector object according to data\n",
    "#         tf_idf_vector_ngram.fit(X_train)\n",
    "        \n",
    "# #         print(dict(zip(tf_idf_vector_ngram.get_feature_names(), tf_idf_vector_ngram.idf_)))\n",
    "# #         print(tf_idf_vector_ngram.get_feature_names())\n",
    "        \n",
    "#         # transform data using tf-idf vector object\n",
    "#         X_train_feature_matrix = tf_idf_vector_ngram.transform(X_train)\n",
    "#         X_test_feature_matrix = tf_idf_vector_ngram.transform(X_test)\n",
    "        return X_train_feature_matrix, X_test_feature_matrix\n",
    "    else:\n",
    "#         # create count vector object \n",
    "        count_vector = CountVectorizer(analyzer='word', ngram_range=(1,ngram), max_df=0.95, max_features=None)\n",
    "#         # fit (create matrix) the count vector object according to data\n",
    "        count_vector.fit(X_train)\n",
    "#         print(count_vector.get_feature_names())\n",
    "#         print(\"Vocabulary content:\\n {}\".format(count_vector.vocabulary_))\n",
    "\n",
    "#         # transform data using count vector object\n",
    "        X_train_feature_matrix = count_vector.transform(X_train)\n",
    "        X_test_feature_matrix = count_vector.transform(X_test)\n",
    "        return X_train_feature_matrix, X_test_feature_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_tf_idf_ngram, X_dev_test_tf_idf_ngram = feature_extraction_and_modeling(X_train, X_dev_test, feature=\"tf-idf\", ngram=2)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<1x4710033 sparse matrix of type '<class 'numpy.float64'>'\n",
       "\twith 46 stored elements in Compressed Sparse Row format>"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_tf_idf_ngram[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<1x4710033 sparse matrix of type '<class 'numpy.float64'>'\n",
       "\twith 27 stored elements in Compressed Sparse Row format>"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_dev_test_tf_idf_ngram[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Naive Bayes model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "def train_model_naive_bayes(X, y):\n",
    "    ros = RandomOverSampler(random_state=None)\n",
    "    X_resampled, y_resampled = ros.fit_resample(X, y)\n",
    "    nb_classifier = MultinomialNB(alpha=1.0, fit_prior=True)\n",
    "    nb_classifier.fit(X_resampled, y_resampled)\n",
    "    return nb_classifier"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Linear SVM model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_model_LinearSVC(X, y):\n",
    "    ros = RandomOverSampler(random_state=None)\n",
    "    X_resampled, y_resampled = ros.fit_resample(X, y)\n",
    "    svm_classifier = LinearSVC(random_state=None)\n",
    "    svm_classifier.fit(X_resampled, y_resampled)\n",
    "    return svm_classifier"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Logistic Regression model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_model_LogisticRegression(X, y):\n",
    "    ros = RandomOverSampler(random_state=None)\n",
    "    X_resampled, y_resampled = ros.fit_resample(X, y)\n",
    "    lr_classifier = LogisticRegression(n_jobs=1, random_state=0, C=1e5, solver='lbfgs', max_iter=10000, multi_class='multinomial')\n",
    "    lr_classifier.fit(X_resampled, y_resampled)\n",
    "    return lr_classifier"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Testing the data with model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "def test_data_with_model(model, X, y):\n",
    "    prediction = model.predict(X)\n",
    "    accuracy = model.score(X, y)\n",
    "    return prediction, accuracy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Function to print confusion matrix and classification report"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "def matrix_and_report(y_test, y_pred, unique_class):\n",
    "    tf_idf_ngram_cm = metrics.confusion_matrix(y_test, y_pred, labels=unique_class)\n",
    "    figure, axis = plt.subplots(figsize=(10,10))\n",
    "    axis.set_title('Confusion matrix')\n",
    "    sns.heatmap(tf_idf_ngram_cm, annot=True, fmt='d', xticklabels=unique_class, yticklabels=unique_class)\n",
    "    plt.ylabel('Actual')\n",
    "    plt.xlabel('Predicted')\n",
    "    plt.show()\n",
    "    print(\"--------------------------------------------------------------------------\")\n",
    "    print(\"Confusion matrix, without normalization\")\n",
    "    print(metrics.confusion_matrix(y_test, y_pred, labels=unique_class))\n",
    "    plt.show()\n",
    "    print(\"--------------------------------------------------------------------------\")\n",
    "    print(\"Classification Report\")\n",
    "    print(metrics.classification_report(y_test, y_pred, labels=unique_class))\n",
    "    print(\"--------------------------------------------------------------------------\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "nb_model = train_model_naive_bayes(X_train_tf_idf_ngram, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MultinomialNB(alpha=1.0, class_prior=None, fit_prior=True)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nb_model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Using Naive Bayes: Test and make prediction of train data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "nb_prediction, nb_accuracy = test_data_with_model(nb_model, X_dev_test_tf_idf_ngram, y_dev_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "# for y_test, y_test_predict in zip(y_dev_test, nb_prediction):\n",
    "#     if y_test != y_test_predict:\n",
    "#         print(y_test + \"\\t\" + y_test_predict)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Using Naive Bayes: Accuracy of predicted train data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Multinomial Naive Bayes classifier accuracy score for test set=0.8708\n"
     ]
    }
   ],
   "source": [
    "print(\"Multinomial Naive Bayes classifier accuracy score for test set=%0.4f\" % nb_accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Multinomial Naive Bayes - Confusion matrix and classification report of train data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--------------------------------------------------------------------------\n",
      "Confusion matrix, without normalization\n",
      "[[1999    0    0    1    0    0    0    0    0    0    0    0    0]\n",
      " [   0 1990    0    0    0    0    9    0    0    0    1    0    0]\n",
      " [   0    0 1870    0  130    0    0    0    0    0    0    0    0]\n",
      " [   0    0    0 1527    1    0    0    0  472    0    0    0    0]\n",
      " [   0    0   91    0 1909    0    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0    0 2000    0    0    0    0    0    0    0]\n",
      " [   0   17    1    0    0    0 1978    0    2    1    1    0    0]\n",
      " [   0    0    0    0    0    0    0 1837    0    1    0   24  138]\n",
      " [   0    0    0  103    0    0    0    0 1897    0    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0  906 1094    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0  896 1104    0    0]\n",
      " [   0    0    0    0    0    0    0   48    0    0    0 1884   68]\n",
      " [   0    0    0    0    0    0    0  102    0    0    0  159 1739]]\n",
      "--------------------------------------------------------------------------\n",
      "Classification Report\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "          sk       1.00      1.00      1.00      2000\n",
      "          my       0.99      0.99      0.99      2000\n",
      "       pt-PT       0.95      0.94      0.94      2000\n",
      "       es-AR       0.94      0.76      0.84      2000\n",
      "       pt-BR       0.94      0.95      0.95      2000\n",
      "          cz       1.00      1.00      1.00      2000\n",
      "          id       1.00      0.99      0.99      2000\n",
      "          hr       0.92      0.92      0.92      2000\n",
      "       es-ES       0.80      0.95      0.87      2000\n",
      "       en-US       0.50      0.45      0.48      2000\n",
      "       en-GB       0.50      0.55      0.53      2000\n",
      "          sr       0.91      0.94      0.93      2000\n",
      "          bs       0.89      0.87      0.88      2000\n",
      "\n",
      "   micro avg       0.87      0.87      0.87     26000\n",
      "   macro avg       0.87      0.87      0.87     26000\n",
      "weighted avg       0.87      0.87      0.87     26000\n",
      "\n",
      "--------------------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "matrix_and_report(y_dev_test, nb_prediction, train_unique_class)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "svm_model = train_model_LinearSVC(X_train_tf_idf_ngram, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,\n",
       "     intercept_scaling=1, loss='squared_hinge', max_iter=1000,\n",
       "     multi_class='ovr', penalty='l2', random_state=None, tol=0.0001,\n",
       "     verbose=0)"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svm_model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Using Linear SVM: Test and make prediction of train data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "svm_prediction, svm_accuracy = test_data_with_model(svm_model, X_dev_test_tf_idf_ngram, y_dev_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Using Linear SVM: Accuracy of predicted train data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Linear SVM classifier accuracy score for test set=0.8746\n"
     ]
    }
   ],
   "source": [
    "print(\"Linear SVM classifier accuracy score for test set=%0.4f\" % svm_accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Linear SVM - Confusion matrix and classification report of train data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--------------------------------------------------------------------------\n",
      "Confusion matrix, without normalization\n",
      "[[2000    0    0    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0 1972    0    0    0    0   26    0    0    1    1    0    0]\n",
      " [   0    0 1879    0  121    0    0    0    0    0    0    0    0]\n",
      " [   0    0    0 1735    0    0    0    0  265    0    0    0    0]\n",
      " [   0    0  109    1 1890    0    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0    0 2000    0    0    0    0    0    0    0]\n",
      " [   0   18    1    0    0    0 1980    0    0    1    0    0    0]\n",
      " [   0    0    0    0    0    0    0 1842    0    1    0   21  136]\n",
      " [   0    0    0  194    0    0    0    0 1806    0    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0  947 1053    0    0]\n",
      " [   1    0    0    0    0    0    0    0    0  938 1061    0    0]\n",
      " [   0    0    0    0    0    0    0   69    0    0    0 1846   85]\n",
      " [   0    0    0    0    0    0    0  113    0    0    0  105 1782]]\n",
      "--------------------------------------------------------------------------\n",
      "Classification Report\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "          sk       1.00      1.00      1.00      2000\n",
      "          my       0.99      0.99      0.99      2000\n",
      "       pt-PT       0.94      0.94      0.94      2000\n",
      "       es-AR       0.90      0.87      0.88      2000\n",
      "       pt-BR       0.94      0.94      0.94      2000\n",
      "          cz       1.00      1.00      1.00      2000\n",
      "          id       0.99      0.99      0.99      2000\n",
      "          hr       0.91      0.92      0.92      2000\n",
      "       es-ES       0.87      0.90      0.89      2000\n",
      "       en-US       0.50      0.47      0.49      2000\n",
      "       en-GB       0.50      0.53      0.52      2000\n",
      "          sr       0.94      0.92      0.93      2000\n",
      "          bs       0.89      0.89      0.89      2000\n",
      "\n",
      "   micro avg       0.87      0.87      0.87     26000\n",
      "   macro avg       0.87      0.87      0.87     26000\n",
      "weighted avg       0.87      0.87      0.87     26000\n",
      "\n",
      "--------------------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "matrix_and_report(y_dev_test, svm_prediction, train_unique_class)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "# lr_model = train_model_LogisticRegression(X_train_tf_idf_ngram, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "# lr_model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Using Logistic Regression: Test and make prediction of train data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "# lr_prediction, lr_accuracy = test_data_with_model(lr_model, X_dev_test_tf_idf_ngram, y_dev_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Using Logistic Regression: Accuracy of predicted train data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "# print(\"Logistic Regression classifier accuracy score for test set=%0.4f\" % lr_accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Logistic Regression - Confusion matrix and classification report of train data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "# matrix_and_report(y_dev_test, lr_prediction, train_unique_class)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Testing on test data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Read file and preprocess the file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_texts, test_classes = read_file_and_preprocess(file_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Začátkem druhé poloviny dvacátého století dominovala až apokalyptická vize o katastrofálním přelidnění světa Robert Mcamara bývalý ministr národní obrany týmu prezidenta JF Kennedyho a posléze prezident Světové banky přirovnal nebezpečí nadcházející populační exploze k neméně ničivému dopadu jaký by způsobila atomová válka Takovým neomalthusiánským pesimismem se zejména vyznačovali ekologové a biologové Opačné optimistické stanovisko se spíš ozývalo z řad ekonomů Ti důvěřovali ve schopnost lidstva se umem důvtipem imaginací přizpůsobit úspěšně reagovat na jakkoliv náročné problémy způsobené přílišným vzrůstem populace Ona tzv zelená revoluce schopnost důkladně zvětšit produkci rýže a tak odvrátit hladomor zejména v Asii je často uváděna jako přesvědčivý příklad\n",
      "cz\n"
     ]
    }
   ],
   "source": [
    "print(test_texts[0])\n",
    "print(test_classes[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Get all unique class labels (unique languages)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_unique_class = unique_class_label(test_classes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['sk', 'pt-PT', 'es-AR', 'my', 'pt-BR', 'cz', 'id', 'hr', 'es-ES', 'en-US', 'en-GB', 'sr', 'bs']\n"
     ]
    }
   ],
   "source": [
    "print(test_unique_class)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### print each line for each language of test set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Začátkem druhé poloviny dvacátého století dominovala až apokalyptická vize o katastrofálním přelidnění světa Robert Mcamara bývalý ministr národní obrany týmu prezidenta JF Kennedyho a posléze prezident Světové banky přirovnal nebezpečí nadcházející populační exploze k neméně ničivému dopadu jaký by způsobila atomová válka Takovým neomalthusiánským pesimismem se zejména vyznačovali ekologové a biologové Opačné optimistické stanovisko se spíš ozývalo z řad ekonomů Ti důvěřovali ve schopnost lidstva se umem důvtipem imaginací přizpůsobit úspěšně reagovat na jakkoliv náročné problémy způsobené přílišným vzrůstem populace Ona tzv zelená revoluce schopnost důkladně zvětšit produkci rýže a tak odvrátit hladomor zejména v Asii je často uváděna jako přesvědčivý příklad\tcz\n",
      "Pošto je ULX obećao da će obaviti istragu o prošlosedmičnom izbijanju nasilja na sjeveru Kosova taj incident predstavlja još jedan ispit kapaciteta misije da doprinese jačanju vladavine prava\tbs\n",
      "De todas as provações que teve de suplantar ao longo da vida qual foi a mais difícil O início Qualquer começo apresenta dificuldades que parecem intransponíveis Mas tive sempre a minha mãe do meu lado Foi ela quem me ajudou a encontrar forças para enfrentar as situações mais decepcionantes negativas as que me punham mesmo furiosa\tpt-PT\n",
      "Al parecer Andrea Guasch pone que una relación a distancia es muy difícil de llevar como excusa Algo con lo que por lo visto Alex Lequio no está nada de acuerdo ¿O es que más bien ya ha conseguido la fama que andaba buscando\tes-ES\n",
      "Vo väčšine golfových rezortov ide o veľký komplex niekoľkých ihrísk blízko pri sebe spojených s hotelmi a ďalšími možnosťami trávenia voľného času – nie vždy sú manželky či deti nadšenými golfistami a tak potrebujú iný druh vyžitia Zaujímavé kombinácie ponúkajú aj rakúske švajčiarske či talianske Alpy kde sa dá v zime lyžovať a v lete hrať golf pod vysokými alpskými končiarmi\tsk\n",
      "Bubnjar Ozgur Can Onej kaže da je vrovizija inspirisala grupu da napiše nove pesme pesme koje kako kaže jedva čekaju da podele sa svetom\tsr\n",
      "O atacante havia dito que tinha esperança de voltar a jogar na competição Mas de acordo com o médico Homero De Agostino um exame realizado na manhã de ontem mostrou que ele não tem mais chances de jogar\tpt-BR\n",
      "Para los empresarios locales contar con el servicio es fundamental para asegurar la competitividad de la economía “s tan importante como el mantenimiento de las rutas que nos unen a Santa Fe a Freyre a Rosario Rafaela es una ciudad del interior no es una capital de provincia está lejos de un río de una autopista entonces los extranjeros que vienen a la ciudad o los porteños que vienen a hacer negocios nos dicen que se tarda menos venir de Dinamarca a Buenos Aires que de Buenos Aires a Rafaela” se quejaba hace sólo una semana atrás Oscar Parra presidente del Centro Comercial e Industrial de Rafaela en declaraciones a una radio local\tes-AR\n",
      "KUALA LUMPUR Keputusan kerajaan untuk tidak menambah bilangan guru sandaran tidak terlatih GSTT adalah bagi memastikan GSTT sedia ada dapat diserap menjadi guru sandaran terlatih GST terlebih dahulu selepas latihan ikhtisas diberikan kepada mereka\tmy\n",
      "Bungsu dari tiga bersaudara buah kasih pasangan di Lamanepa dan Widiany ini mengatakan jumlah penumpang sakit yang ditanganinnya tidak banyak seperti umumnya di rumah sakit atau puskesmas amun yang dilihat bukan seberapa banyak atau sedikitnya melainkan pelayanan medis yang diberikan harus membuat pasien nyaman\tid\n",
      "Srbija i Crna Gora bila je treća po broju turista koji su posjetili Bugarsku  godine kada je skoro  njezinih građana odlučilo provesti svoj odmor u toj zemlji\thr\n",
      "AUBUR HILLS  In the first official meaningless game of the season the Pistons played as if their lives depended on it\ten-US\n",
      "arlier that day clad in a protective white suit Paul  a special enforcement officer from the nvironment Agency  had painstakingly examined the entire field and found hundreds of pieces of asbestos no bigger than a golf ball\ten-GB\n"
     ]
    }
   ],
   "source": [
    "print_line_each_language(test_texts, test_classes)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Feature selection and modeling\n",
    "Count - bag of words (word)\n",
    "<br>\n",
    "tf-idf - unigram, bigram (ngram)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train_tf_idf_ngram, test_tf_idf_ngram = feature_extraction_and_modeling(X_train, test_texts, feature=\"tf-idf\", ngram=2)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<1x4710033 sparse matrix of type '<class 'numpy.float64'>'\n",
       "\twith 94 stored elements in Compressed Sparse Row format>"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_tf_idf_ngram[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Using Naive Bayes: Test and make prediction of test data "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "nb_test_prediction, nb_test_accuracy = test_data_with_model(nb_model, test_tf_idf_ngram, test_classes)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Using Naive Bayes:  Accuracy of predicted test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Multinomial Naive Bayes classifier accuracy score for test set=  87.45238095238095\n"
     ]
    }
   ],
   "source": [
    "print(\"Multinomial Naive Bayes classifier accuracy score for test set= \", nb_test_accuracy*100)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Using Naive Bayes: Print Confusion matrix and classification report of test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--------------------------------------------------------------------------\n",
      "Confusion matrix, without normalization\n",
      "[[1000    0    0    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0  933    0    0   67    0    0    0    0    0    0    0    0]\n",
      " [   0    0  764    0    0    0    0    0  236    0    0    0    0]\n",
      " [   0    0    0  995    0    0    4    0    0    0    1    0    0]\n",
      " [   0   52    0    0  948    0    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0    0  999    0    0    0    0    1    0    0]\n",
      " [   0    0    0    7    0    0  991    0    0    1    1    0    0]\n",
      " [   0    0    0    0    0    0    0  903    0    0    1   14   82]\n",
      " [   0    0   49    0    0    0    0    0  951    0    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0  356  444    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0  442  358    0    0]\n",
      " [   0    0    0    0    0    0    0   15    0    0    0  934   51]\n",
      " [   0    0    0    0    0    0    0   34    0    0    0   79  887]]\n",
      "--------------------------------------------------------------------------\n",
      "Classification Report\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "          sk       1.00      1.00      1.00      1000\n",
      "       pt-PT       0.95      0.93      0.94      1000\n",
      "       es-AR       0.94      0.76      0.84      1000\n",
      "          my       0.99      0.99      0.99      1000\n",
      "       pt-BR       0.93      0.95      0.94      1000\n",
      "          cz       1.00      1.00      1.00      1000\n",
      "          id       1.00      0.99      0.99      1000\n",
      "          hr       0.95      0.90      0.93      1000\n",
      "       es-ES       0.80      0.95      0.87      1000\n",
      "       en-US       0.45      0.45      0.45       800\n",
      "       en-GB       0.44      0.45      0.45       800\n",
      "          sr       0.91      0.93      0.92      1000\n",
      "          bs       0.87      0.89      0.88      1000\n",
      "\n",
      "   micro avg       0.87      0.87      0.87     12600\n",
      "   macro avg       0.86      0.86      0.86     12600\n",
      "weighted avg       0.88      0.87      0.87     12600\n",
      "\n",
      "--------------------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "matrix_and_report(test_classes, nb_test_prediction, test_unique_class)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Not matched sentences"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "# for sent, y_test, y_test_predict in zip(test_texts, test_classes, nb_test_prediction):\n",
    "#     if y_test==\"pt-PT\" and y_test_predict==\"es-ES\":\n",
    "#         print(sent + \"\\t\" + y_test + \"\\t\" + y_test_predict)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Using Linear SVM: Test and make prediction of test data "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "svm_test_prediction, svm_test_accuracy = test_data_with_model(svm_model, test_tf_idf_ngram, test_classes)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Using Linear SVM:  Accuracy of predicted test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Linear SVM classifier accuracy score for test set=0.8806\n"
     ]
    }
   ],
   "source": [
    "print(\"Linear SVM classifier accuracy score for test set=%0.4f\" % svm_test_accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Using Linear SVM:  Print Confusion matrix and classification report of test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--------------------------------------------------------------------------\n",
      "Confusion matrix, without normalization\n",
      "[[1000    0    0    0    0    0    0    0    0    0    0    0    0]\n",
      " [   0  931    0    0   69    0    0    0    0    0    0    0    0]\n",
      " [   1    0  873    0    0    0    0    0  126    0    0    0    0]\n",
      " [   0    0    0  990    0    0   10    0    0    0    0    0    0]\n",
      " [   0   65    1    0  934    0    0    0    0    0    0    0    0]\n",
      " [   0    0    0    0    0  999    0    0    0    0    1    0    0]\n",
      " [   0    0    0   11    0    0  988    0    0    0    1    0    0]\n",
      " [   0    0    0    0    0    0    0  897    0    0    0   17   86]\n",
      " [   0    1   76    0    0    0    0    0  923    0    0    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0  376  424    0    0]\n",
      " [   0    0    0    0    0    0    0    0    0  426  374    0    0]\n",
      " [   0    0    0    0    0    0    0   23    0    0    0  919   58]\n",
      " [   0    0    0    0    0    0    0   44    0    0    0   65  891]]\n",
      "--------------------------------------------------------------------------\n",
      "Classification Report\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "          sk       1.00      1.00      1.00      1000\n",
      "       pt-PT       0.93      0.93      0.93      1000\n",
      "       es-AR       0.92      0.87      0.90      1000\n",
      "          my       0.99      0.99      0.99      1000\n",
      "       pt-BR       0.93      0.93      0.93      1000\n",
      "          cz       1.00      1.00      1.00      1000\n",
      "          id       0.99      0.99      0.99      1000\n",
      "          hr       0.93      0.90      0.91      1000\n",
      "       es-ES       0.88      0.92      0.90      1000\n",
      "       en-US       0.47      0.47      0.47       800\n",
      "       en-GB       0.47      0.47      0.47       800\n",
      "          sr       0.92      0.92      0.92      1000\n",
      "          bs       0.86      0.89      0.88      1000\n",
      "\n",
      "   micro avg       0.88      0.88      0.88     12600\n",
      "   macro avg       0.87      0.87      0.87     12600\n",
      "weighted avg       0.88      0.88      0.88     12600\n",
      "\n",
      "--------------------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "matrix_and_report(test_classes, svm_test_prediction, test_unique_class)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Not matched sentences"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "# for sent, y_test, y_test_predict in zip(test_texts, test_classes, svm_test_prediction):\n",
    "#     if y_test != y_test_predict:\n",
    "#         print(sent + \"\\t\" + y_test + \"\\t\" + y_test_predict)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Using Logistic Regression: Test and make prediction of test data "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "# lr_test_prediction, lr_test_accuracy = test_data_with_model(lr_model, test_tf_idf_ngram, test_classes)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Using Logistic Regression:  Accuracy of predicted test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "# print(\"Logistic Regression classifier accuracy score for test set=%0.4f\" % lr_test_accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Using Logistic Regression:  Print Confusion matrix and classification report of test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "# matrix_and_report(test_classes, lr_test_prediction, test_unique_class)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}