NLP_Challenge_2020_start.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "\n",
    "from sklearn.metrics import accuracy_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# on lit avec pandas les fichiers de train et de test\n",
    "train =pd.read_pickle('train.pkl')\n",
    "test = pd.read_pickle('test.pkl')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# on vectorise les textes des jeux de train et de test\n",
    "count_vector = CountVectorizer()\n",
    "X_train_count = count_vector.fit_transform(train.X)\n",
    "X_test_count = count_vector.transform(test.X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Entrainement d'un KNN et prediction sur la vectorisation du jeu de test\n",
    "neigh = KNeighborsClassifier(n_neighbors=10)\n",
    "neigh.fit(X_train_count, train.y)\n",
    "y_pred_train = neigh.predict(X_train_count)\n",
    "y_pred_test = neigh.predict(X_test_count)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# on exporte les predictions sur le jeu de test\n",
    "pd.DataFrame(y_pred_test).to_csv('y_pred_test.csv' )"
   ]
  }
 ],
 "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": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}