test.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "1ce0c6439adb27d7",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "# Test of flagman performance\n",
    "\n",
    "In this notebook we compare the improvement made by the hybrid model over the baseline."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "initial_id",
   "metadata": {
    "collapsed": true,
    "ExecuteTime": {
     "end_time": "2024-05-31T12:13:03.103626600Z",
     "start_time": "2024-05-31T12:12:57.698453800Z"
    }
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "from movie_lens_lib import *\n",
    "from visualization import hybrid_vs_baselines_visualization"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fc6fb4e2a4c5bcf5",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "## Constants"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "32e75f71c8bee821",
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-05-31T12:13:03.119826200Z",
     "start_time": "2024-05-31T12:13:03.103626600Z"
    }
   },
   "outputs": [],
   "source": [
    "n_movie_clusters = 5\n",
    "rating_multiplier = 5\n",
    "year_multiplier = 0.05\n",
    "weight_genre, weight_cluster, weight_movie = 0.35, 0.45, 0.2\n",
    "train_size = 0.9\n",
    "random_state = 42"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5c86840d37e286e3",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "## Import & Split dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "6735412f0824a406",
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-05-31T12:13:39.566262500Z",
     "start_time": "2024-05-31T12:13:03.111289900Z"
    }
   },
   "outputs": [],
   "source": [
    "ratings_df = pd.read_csv(\"data/ratings.csv\")\n",
    "movies_df = pd.read_csv(\"data/movies.csv\", index_col=\"movieId\")\n",
    "\n",
    "X = ratings_df.drop([\"rating\"], axis=1)\n",
    "y = ratings_df[\"rating\"]\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, train_size=train_size, random_state=random_state)\n",
    "ratings_train_df = pd.concat([X_train, y_train], axis=1)\n",
    "ratings_test_df = pd.concat([X_test, y_test], axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d467359b2c2d5e42",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "## Preprocess"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "34c1f982e01c9bdc",
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-05-31T12:13:42.368095300Z",
     "start_time": "2024-05-31T12:13:39.562259900Z"
    }
   },
   "outputs": [],
   "source": [
    "movies_hot_df = PreProcessingAggregated().transform((movies_df, ratings_train_df))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3a0e9ce4a12a3bb5",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "## Baseline performance\n",
    "We are going to compare 3 baseline models in this section.\n",
    "\n",
    "### Naive constant\n",
    "\n",
    "As the first baseline we predict rating of **2.5** for every pair of USER - MOVIE."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "89852ef664dce374",
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-05-31T12:13:42.613422100Z",
     "start_time": "2024-05-31T12:13:42.373095Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MSE: 2.221\n",
      "MAE: 1.297\n",
      "ACCURACY: 0.453\n"
     ]
    }
   ],
   "source": [
    "y_pred = np.zeros(y_test.shape[0]) + 2.5\n",
    "print_stats(naive_baseline_results := get_performance_stats(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "88bedbb5dd552d7a",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "### Mean rating over the training set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "c6b26484506f9e7e",
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-05-31T12:13:42.980227400Z",
     "start_time": "2024-05-31T12:13:42.613422100Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MSE: 1.133\n",
      "MAE: 0.843\n",
      "ACCURACY: 0.675\n"
     ]
    }
   ],
   "source": [
    "y_pred = np.zeros(y_test.shape[0]) + y_train.mean()\n",
    "print_stats(global_mean_baseline_results := get_performance_stats(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e1821c5b235e7af5",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "### Mean rating for the movie\n",
    "\n",
    "We have implemented the regressor for this task, and it is put into movie_lens_lib.py"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "d6a883502bc819be",
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-05-31T12:13:43.607736400Z",
     "start_time": "2024-05-31T12:13:42.978225Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MSE: 0.932\n",
      "MAE: 0.745\n",
      "ACCURACY: 0.727\n"
     ]
    }
   ],
   "source": [
    "movie_regressor = MovieBasedRegressor().fit(movies_hot_df)\n",
    "movie_predictions = movie_regressor.predict(X_test, False)\n",
    "print_stats(movie_regressor_results := get_performance_stats(y_test, movie_predictions))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e086dad2e9c83d37",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "## Hybrid performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "b9481c48cff9490d",
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-05-31T12:16:32.092924800Z",
     "start_time": "2024-05-31T12:13:43.606737100Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MSE: 0.75\n",
      "MAE: 0.664\n",
      "ACCURACY: 0.78\n"
     ]
    }
   ],
   "source": [
    "hybrid_regressor = HybridRegressor(\n",
    "    movies_hot_df,\n",
    "    cluster_based_regressor = ClusterBasedRegressor(\n",
    "        movies_hot_df,\n",
    "        n_movie_clusters,\n",
    "        rating_multiplier,\n",
    "        year_multiplier,\n",
    "        random_state\n",
    "    )\n",
    ").fit(X_train, y_train)\n",
    "\n",
    "hybrid_predictions = hybrid_regressor.predict(X_test, False)\n",
    "print_stats(hybrid_results := get_performance_stats(y_test, hybrid_predictions))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7fadc28bf596897e",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "## Performance comparison"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "e821a15051e8e1d6",
   "metadata": {
    "collapsed": false,
    "ExecuteTime": {
     "end_time": "2024-05-31T12:16:32.794326300Z",
     "start_time": "2024-05-31T12:16:32.081902100Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 1200x500 with 3 Axes>",
      "image/png": "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"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "results = {\n",
    "    \"Naive Baseline\": naive_baseline_results,\n",
    "    \"Global Mean Baseline\": global_mean_baseline_results,\n",
    "    \"Movie Regressor\": movie_regressor_results,\n",
    "    \"Hybrid\": hybrid_results\n",
    "}\n",
    "hybrid_vs_baselines_visualization(results)"
   ]
  }
 ],
 "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": "3.11.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}