{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import csv\n",
    "from gensim.models import Word2Vec\n",
    "from sklearn.metrics import accuracy_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_documents = []\n",
    "train_classes = []\n",
    "\n",
    "with open('train/train.tsv', 'r', encoding='utf-8') as file:\n",
    "    lines = file.readlines()\n",
    "    for line in lines:\n",
    "        elements = line.split('\\t')\n",
    "        train_classes.append(int(elements[0]))\n",
    "        train_documents.append(elements[1].lower())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Word2Vec(sentences=[doc.split() for doc in train_documents], vector_size=100, window=5, min_count=1, workers=4)\n",
    "model.save(\"word2vec.model\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_test_data(path):\n",
    "    with open(path, 'r', encoding='utf-8') as file:\n",
    "        test_data = []\n",
    "        lines = file.readlines()\n",
    "        for line in lines:\n",
    "            test_data.append(line.strip().split('\\t')[0])\n",
    "        return test_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "dev0_documents = [x.lower() for x in get_test_data('dev-0/in.tsv')]\n",
    "dev0_classes = [int(x) for x in get_test_data('dev-0/expected.tsv')]\n",
    "a_documents = [x.lower() for x in get_test_data('test-A/in.tsv')]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def document_to_word2vec(doc):\n",
    "    vector = np.zeros(model.vector_size)\n",
    "    words = doc.split()\n",
    "    words_present = 0\n",
    "    for word in words:\n",
    "        if word in model.wv:\n",
    "            word_vector = model.wv.get_vector(word)\n",
    "            vector += word_vector\n",
    "            words_present += 1\n",
    "    if words_present > 0:\n",
    "        vector = vector / words_present\n",
    "    return vector"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_loaded = Word2Vec.load(\"word2vec.model\")\n",
    "\n",
    "train_documents_word2vec = [document_to_word2vec(doc) for doc in train_documents]\n",
    "dev0_documents_word2vec = [document_to_word2vec(doc) for doc in dev0_documents]\n",
    "a_documents_word2vec = [document_to_word2vec(doc) for doc in a_documents]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.gaussian_process.kernels import RBF\n",
    "from sklearn.pipeline import make_pipeline\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "from sklearn.neural_network import MLPClassifier\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.naive_bayes import GaussianNB\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.discriminant_analysis import QuadraticDiscriminantAnalysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test accuracy for classifier Linear SVM: 0.9745047688921497\n",
      "Test accuracy for classifier Naive Bayes: 0.892516507703595\n",
      "Test accuracy for classifier Random Forest: 0.960564930300807\n",
      "Test accuracy for classifier QDA: 0.923881144534116\n"
     ]
    }
   ],
   "source": [
    "names = [\n",
    "    \"Linear SVM\",\n",
    "    #\"Neural Net\",\n",
    "    \"Naive Bayes\",\n",
    "    \"Random Forest\",\n",
    "    \"QDA\"\n",
    "]\n",
    "\n",
    "classifiers = [\n",
    "    MLPClassifier(alpha=1, max_iter=1000, random_state=42),\n",
    "    #SVC(gamma=2, C=1, random_state=42),\n",
    "    GaussianNB(),\n",
    "    RandomForestClassifier(),\n",
    "    QuadraticDiscriminantAnalysis()\n",
    "]\n",
    "\n",
    "best_accuracy = 0\n",
    "best_classifier_name = \"\"\n",
    "\n",
    "for name, clf in zip(names, classifiers):\n",
    "    clf = make_pipeline(StandardScaler(), clf)\n",
    "    clf.fit(train_documents_word2vec, train_classes)\n",
    "\n",
    "    dev0_predictions = clf.predict(dev0_documents_word2vec)\n",
    "    a_predictions = clf.predict(a_documents_word2vec)\n",
    "    dev0_accuracy = accuracy_score(dev0_classes, dev0_predictions)\n",
    "    print(\"Test accuracy for classifier \" + name + \":\", dev0_accuracy)\n",
    "\n",
    "    if dev0_accuracy > best_accuracy:\n",
    "        best_accuracy = dev0_accuracy\n",
    "        best_classifier_name = name\n",
    "        best_dev0_predictions = dev0_predictions\n",
    "        best_a_predictions = a_predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('dev-0/out.tsv', 'w+', newline='', encoding='utf-8') as file:\n",
    "    writer = csv.writer(file, delimiter='\\t')\n",
    "    for prediction in best_dev0_predictions:\n",
    "        writer.writerow([prediction])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('test-A/out.tsv', 'w+', newline='', encoding='utf-8') as file:\n",
    "    writer = csv.writer(file, delimiter='\\t')\n",
    "    for prediction in best_a_predictions:\n",
    "        writer.writerow([prediction])"
   ]
  }
 ],
 "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": "3.9.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}