12-collaborative_filtering_deep.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Collaborative filtering deep"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from functools import reduce\n",
    "import pickle\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import QuantileTransformer\n",
    "from sklearn.metrics import mean_squared_error, mean_absolute_error\n",
    "from sklearn.metrics import classification_report, roc_curve, auc, confusion_matrix\n",
    "from sklearn.utils.class_weight import compute_sample_weight, compute_class_weight\n",
    "\n",
    "from tensorflow.keras import Input\n",
    "from tensorflow.keras.layers import Flatten, Dot, Concatenate\n",
    "from tensorflow.keras.layers import Embedding, SpatialDropout1D, Dense, Dropout, BatchNormalization\n",
    "from tensorflow.keras.models import load_model, Model\n",
    "from tensorflow.keras.callbacks import EarlyStopping\n",
    "from tensorflow.keras.utils import to_categorical, plot_model\n",
    "\n",
    "pd.set_option('display.max_columns', None)\n",
    "\n",
    "def save(obj, path):\n",
    "    with open(path, 'bw') as file:\n",
    "        pickle.dump(obj, file)\n",
    "\n",
    "def load(path):\n",
    "    with open(path, 'br') as file:\n",
    "        return pickle.load(file)\n",
    "\n",
    "def missing(df):\n",
    "    return df[df.isna().any(axis='columns')]\n",
    "\n",
    "def chain_layers(*layers):\n",
    "    return reduce(lambda output, layer: layer(output), layers)\n",
    "\n",
    "def plot_loss(history):\n",
    "    n_epoch = len(history['loss'])\n",
    "    xs = np.arange(1, n_epoch + 1)\n",
    "    fig, ax = plt.subplots()\n",
    "    ax.set_ylabel('Loss')\n",
    "    ax.set_xlabel('Epoch')\n",
    "    ax.plot(xs, history['loss'], label='Train')\n",
    "    ax.plot(xs, history['val_loss'], label='Validation')\n",
    "    ax.legend()\n",
    "    return fig\n",
    "\n",
    "def plot_acc(history):\n",
    "    n_epoch = len(history['accuracy'])\n",
    "    xs = np.arange(1, n_epoch + 1)\n",
    "    fig, [loss, acc] = plt.subplots(2, 1, sharex=True)\n",
    "    fig.subplots_adjust(hspace=0)\n",
    "    loss.set_ylabel('Loss')\n",
    "    loss.plot(xs, history['loss'], label='Train')\n",
    "    loss.plot(xs, history['val_loss'], label='Validation')\n",
    "    acc.set_xlabel('Epoch')\n",
    "    acc.set_ylabel('Accuracy')\n",
    "    acc.plot(xs, history['accuracy'])\n",
    "    acc.plot(xs, history['val_accuracy'])\n",
    "    fig.legend()\n",
    "    return fig\n",
    "\n",
    "def print_rg_report(y, y_pred):\n",
    "    print(f'MSE = {mean_squared_error(y, y_pred)}\\nMAE = {mean_absolute_error(y, y_pred)}')\n",
    "\n",
    "def plot_preds(y, y_pred):\n",
    "    n_classes = len(np.unique(y))\n",
    "    fig, ax = plt.subplots(figsize=(n_classes, n_classes))\n",
    "    ax.set_xlim(0, n_classes + 1)\n",
    "    ax.set_xticks(np.arange(1, n_classes + 1))\n",
    "    ax.set_xlabel('Actual')\n",
    "    ax.set_ylabel('Predicted')\n",
    "    ax.scatter(y, y_pred, s=64, alpha=1 / 256)\n",
    "    return fig\n",
    "\n",
    "def print_cl_report(y, y_pred):\n",
    "    print(classification_report(y, y_pred))\n",
    "\n",
    "def plot_rocc(y, Y_prob):\n",
    "    n_classes = Y_prob.shape[1]\n",
    "    first_class = (n_classes == 2 and 1) or 0\n",
    "    fig, ax = plt.subplots()\n",
    "    ax.set_xlim(0, 1)\n",
    "    ax.set_ylim(0, 1)\n",
    "    ax.set_xlabel('FPR')\n",
    "    ax.set_ylabel('TPR')\n",
    "    ax.plot([0, 1], [0, 1], \"--\")\n",
    "    for cl in range(first_class, n_classes):\n",
    "        fpr, tpr, _ths = roc_curve(y, Y_prob[:, cl], pos_label=cl)\n",
    "        ax.plot(fpr, tpr, label=f'{cl}, AUC = {auc(fpr, tpr):.2f}')\n",
    "    ax.legend()\n",
    "    return fig\n",
    "\n",
    "def plot_confmat(y, y_pred):\n",
    "    n_classes = len(np.unique(y))\n",
    "    confmat = confusion_matrix(y, y_pred)\n",
    "    max_v = confmat.max()\n",
    "    fig, ax = plt.subplots(figsize=(n_classes, n_classes))\n",
    "    xys = np.arange(n_classes)\n",
    "    ax.set_xticks(xys)\n",
    "    ax.set_yticks(xys)\n",
    "    ax.set_xlabel('Predicted')\n",
    "    ax.set_ylabel('Actual')\n",
    "    ax.imshow(confmat, vmin=max_v * -1, vmax=max_v, cmap='coolwarm')\n",
    "    for actual in range(n_classes):\n",
    "        for predicted in range(n_classes):\n",
    "            ax.text(predicted, actual, confmat[actual, predicted], ha='center', va='center')\n",
    "    return fig"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "movies = pd.read_parquet('data/movies_active.parquet')\n",
    "ratings = pd.read_parquet('data/ratings_active.parquet')\n",
    "users = pd.read_parquet('data/users_active.parquet')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "n_classes = ratings['rating'].nunique()\n",
    "user_dim, user_voc_size = (1,), users['user_id'].max()\n",
    "movie_dim, movie_voc_size = (1,), movies['movie_id'].max()\n",
    "X_train, X_test, y_train, y_test, Y_train = load('data/train_test_active.pickle')[3:]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Regressie"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "emb_size = 16\n",
    "\n",
    "user_input = Input(user_dim, name='user_id')\n",
    "user_emb = chain_layers(\n",
    "    user_input,\n",
    "    Embedding(user_voc_size + 1, emb_size, name='user_emb'),\n",
    "    SpatialDropout1D(0.1),\n",
    "    Flatten()\n",
    ")\n",
    "movie_input = Input(movie_dim, name='movie_id')\n",
    "movie_emb = chain_layers(\n",
    "    movie_input,\n",
    "    Embedding(movie_voc_size + 1, emb_size, name='movie_emb'),\n",
    "    SpatialDropout1D(0.1),\n",
    "    Flatten()\n",
    ")\n",
    "user_movie = chain_layers(\n",
    "    Concatenate()([user_emb, movie_emb]),\n",
    "    Dense(64, 'relu'),\n",
    "    Dropout(0.1),\n",
    "    Dense(64, 'relu'),\n",
    "    Dropout(0.1),\n",
    "    Dense(1)\n",
    ")\n",
    "movies_cf_r_d = Model([user_input, movie_input], user_movie, name='movies_cf_r_d')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"movies_cf_r_d\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "user_id (InputLayer)            [(None, 1)]          0                                            \n",
      "__________________________________________________________________________________________________\n",
      "movie_id (InputLayer)           [(None, 1)]          0                                            \n",
      "__________________________________________________________________________________________________\n",
      "user_emb (Embedding)            (None, 1, 16)        113968      user_id[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "movie_emb (Embedding)           (None, 1, 16)        526688      movie_id[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "spatial_dropout1d (SpatialDropo (None, 1, 16)        0           user_emb[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "spatial_dropout1d_1 (SpatialDro (None, 1, 16)        0           movie_emb[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "flatten (Flatten)               (None, 16)           0           spatial_dropout1d[0][0]          \n",
      "__________________________________________________________________________________________________\n",
      "flatten_1 (Flatten)             (None, 16)           0           spatial_dropout1d_1[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "concatenate (Concatenate)       (None, 32)           0           flatten[0][0]                    \n",
      "                                                                 flatten_1[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "dense (Dense)                   (None, 64)           2112        concatenate[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "dropout (Dropout)               (None, 64)           0           dense[0][0]                      \n",
      "__________________________________________________________________________________________________\n",
      "dense_1 (Dense)                 (None, 64)           4160        dropout[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "dropout_1 (Dropout)             (None, 64)           0           dense_1[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "dense_2 (Dense)                 (None, 1)            65          dropout_1[0][0]                  \n",
      "==================================================================================================\n",
      "Total params: 646,993\n",
      "Trainable params: 646,993\n",
      "Non-trainable params: 0\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "movies_cf_r_d.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "plot_model(movies_cf_r_d, '/tmp/model.png', rankdir='LR')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/8\n",
      "3011/3011 [==============================] - 42s 14ms/step - loss: 0.7524 - val_loss: 0.4526\n",
      "Epoch 2/8\n",
      "3011/3011 [==============================] - 45s 15ms/step - loss: 0.4789 - val_loss: 0.4421\n",
      "Epoch 3/8\n",
      "3011/3011 [==============================] - 44s 15ms/step - loss: 0.4493 - val_loss: 0.4384\n",
      "Epoch 4/8\n",
      "3011/3011 [==============================] - 46s 15ms/step - loss: 0.4272 - val_loss: 0.4348\n"
     ]
    }
   ],
   "source": [
    "movies_cf_r_d.compile('adam', 'mse')\n",
    "movies_cf_r_d_hist = movies_cf_r_d.fit(\n",
    "    [X_train['user_id'], X_train['movie_id']], y_train, batch_size=128, epochs=8, validation_split=0.2,\n",
    "    callbacks=[EarlyStopping(min_delta=0.1, patience=3, restore_best_weights=True)]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "movies_cf_r_d.save('data/movies_cf_r_d.h5')\n",
    "save(movies_cf_r_d_hist.history, 'data/movies_cf_r_d_history.pickle')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Evalueren"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "movies_cf_r_d = load_model('data/movies_cf_r_d.h5')\n",
    "movies_cf_r_d_history = load('data/movies_cf_r_d_history.pickle')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_loss(movies_cf_r_d_history);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_movies_cf_r_d_test = movies_cf_r_d.predict([X_test['user_id'], X_test['movie_id']])\n",
    "y_movies_cf_r_d_train = movies_cf_r_d.predict([X_train['user_id'], X_train['movie_id']])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test:\n",
      "MSE = 0.4531018225905073\n",
      "MAE = 0.5150060997103595\n",
      "Train:\n",
      "MSE = 0.42221901525166805\n",
      "MAE = 0.4953185273379742\n"
     ]
    }
   ],
   "source": [
    "print('Test:')\n",
    "print_rg_report(y_test, y_movies_cf_r_d_test)\n",
    "print('Train:')\n",
    "print_rg_report(y_train, y_movies_cf_r_d_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_preds(y_test, y_movies_cf_r_d_test);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Classificatie"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "emb_size = 16\n",
    "\n",
    "user_input = Input(user_dim, name='user_id')\n",
    "user_emb = chain_layers(\n",
    "    user_input,\n",
    "    Embedding(user_voc_size + 1, emb_size, name='user_emb'),\n",
    "    SpatialDropout1D(0.1),\n",
    "    Flatten()\n",
    ")\n",
    "movie_input = Input(movie_dim, name='movie_id')\n",
    "movie_emb = chain_layers(\n",
    "    movie_input,\n",
    "    Embedding(movie_voc_size + 1, emb_size, name='movie_emb'),\n",
    "    SpatialDropout1D(0.1),\n",
    "    Flatten()\n",
    ")\n",
    "user_movie = chain_layers(\n",
    "    Concatenate()([user_emb, movie_emb]),\n",
    "    Dense(64, 'relu'),\n",
    "    Dropout(0.1),\n",
    "    Dense(64, 'relu'),\n",
    "    Dropout(0.1),\n",
    "    Dense(n_classes, 'softmax')\n",
    ")\n",
    "movies_cf_c_d = Model([user_input, movie_input], user_movie, name='movies_cf_c_d')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"movies_cf_c_d\"\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "user_id (InputLayer)            [(None, 1)]          0                                            \n",
      "__________________________________________________________________________________________________\n",
      "movie_id (InputLayer)           [(None, 1)]          0                                            \n",
      "__________________________________________________________________________________________________\n",
      "user_emb (Embedding)            (None, 1, 16)        113968      user_id[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "movie_emb (Embedding)           (None, 1, 16)        526688      movie_id[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "spatial_dropout1d_2 (SpatialDro (None, 1, 16)        0           user_emb[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "spatial_dropout1d_3 (SpatialDro (None, 1, 16)        0           movie_emb[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "flatten_2 (Flatten)             (None, 16)           0           spatial_dropout1d_2[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "flatten_3 (Flatten)             (None, 16)           0           spatial_dropout1d_3[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_1 (Concatenate)     (None, 32)           0           flatten_2[0][0]                  \n",
      "                                                                 flatten_3[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "dense_3 (Dense)                 (None, 64)           2112        concatenate_1[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "dropout_2 (Dropout)             (None, 64)           0           dense_3[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "dense_4 (Dense)                 (None, 64)           4160        dropout_2[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "dropout_3 (Dropout)             (None, 64)           0           dense_4[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "dense_5 (Dense)                 (None, 5)            325         dropout_3[0][0]                  \n",
      "==================================================================================================\n",
      "Total params: 647,253\n",
      "Trainable params: 647,253\n",
      "Non-trainable params: 0\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "movies_cf_c_d.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "plot_model(movies_cf_c_d, '/tmp/model.png', rankdir='LR')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/8\n",
      "3011/3011 [==============================] - 45s 15ms/step - loss: 0.9830 - accuracy: 0.5710 - val_loss: 0.9157 - val_accuracy: 0.5992\n",
      "Epoch 2/8\n",
      "3011/3011 [==============================] - 45s 15ms/step - loss: 0.8829 - accuracy: 0.6180 - val_loss: 0.9045 - val_accuracy: 0.6028\n",
      "Epoch 3/8\n",
      "3011/3011 [==============================] - 45s 15ms/step - loss: 0.8468 - accuracy: 0.6340 - val_loss: 0.9083 - val_accuracy: 0.6016\n",
      "Epoch 4/8\n",
      "3011/3011 [==============================] - 43s 14ms/step - loss: 0.8224 - accuracy: 0.6440 - val_loss: 0.9137 - val_accuracy: 0.6000\n"
     ]
    }
   ],
   "source": [
    "movies_cf_c_d.compile('adam', 'categorical_crossentropy', ['accuracy'])\n",
    "movies_cf_c_d_hist = movies_cf_c_d.fit(\n",
    "    [X_train['user_id'], X_train['movie_id']], Y_train, batch_size=128, epochs=8, validation_split=0.2,\n",
    "    callbacks=[EarlyStopping('val_accuracy', min_delta=0.005, patience=3, restore_best_weights=True)]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "movies_cf_c_d.save('data/movies_cf_c_d.h5')\n",
    "save(movies_cf_c_d_hist.history, 'data/movies_cf_c_d_history.pickle')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Evalueren"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "movies_cf_c_d = load_model('data/movies_cf_c_d.h5')\n",
    "movies_cf_c_d_history = load('data/movies_cf_c_d_history.pickle')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_acc(movies_cf_c_d_history);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "Y_movies_cf_c_d_test = movies_cf_c_d.predict([X_test['user_id'], X_test['movie_id']])\n",
    "y_movies_cf_c_d_test = Y_movies_cf_c_d_test.argmax(axis=1)\n",
    "Y_movies_cf_c_d_train = movies_cf_c_d.predict([X_train['user_id'], X_train['movie_id']])\n",
    "y_movies_cf_c_d_train = Y_movies_cf_c_d_train.argmax(axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test:\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/nicolas/.local/lib/python3.8/site-packages/sklearn/metrics/_classification.py:1272: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.00      0.00      0.00      1646\n",
      "           1       0.00      0.00      0.00      4616\n",
      "           2       0.57      0.49      0.52     42436\n",
      "           3       0.60      0.75      0.67     77496\n",
      "           4       0.63      0.50      0.55     34373\n",
      "\n",
      "    accuracy                           0.60    160567\n",
      "   macro avg       0.36      0.35      0.35    160567\n",
      "weighted avg       0.58      0.60      0.58    160567\n",
      "\n",
      "Train:\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.00      0.00      0.00      4940\n",
      "           1       0.00      0.00      0.00     13846\n",
      "           2       0.60      0.52      0.56    127306\n",
      "           3       0.63      0.78      0.70    232487\n",
      "           4       0.66      0.53      0.59    103119\n",
      "\n",
      "    accuracy                           0.63    481698\n",
      "   macro avg       0.38      0.37      0.37    481698\n",
      "weighted avg       0.60      0.63      0.61    481698\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print('Test:')\n",
    "print_cl_report(y_test - 1, y_movies_cf_c_d_test)\n",
    "print('Train:')\n",
    "print_cl_report(y_train - 1, y_movies_cf_c_d_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_rocc(y_test - 1, Y_movies_cf_c_d_test);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_confmat(y_test - 1, y_movies_cf_c_d_test);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MSE = 0.5476031812265285\n",
      "MAE = 0.4438085036153132\n"
     ]
    }
   ],
   "source": [
    "print_rg_report(y_test - 1, y_movies_cf_c_d_test)"
   ]
  }
 ],
 "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": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}