3_Boosting.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd \n",
    "import numpy as np \n",
    "from math import sqrt\n",
    "\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from catboost import CatBoostRegressor\n",
    "\n",
    "from sklearn.preprocessing import  StandardScaler, OneHotEncoder\n",
    "from sklearn.metrics import mean_absolute_error, mean_squared_error\n",
    "from sklearn.model_selection import train_test_split, cross_val_score\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.set_option('display.max_columns', None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('dataset.csv',index_col=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Prepare data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.drop(['date_install','first_command_time','birthday','sex','time_confirm_email'],axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "int_cols = df.select_dtypes(include=['int', 'float']).columns\n",
    "int_cols = int_cols.drop('target_game_currency')\n",
    "obj_cols = ['country','network_name']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def prepare_obj_data(df, obj_cols):\n",
    "    '''\n",
    "    Prepare data with type object\n",
    "    \n",
    "    Args:\n",
    "        df: pd.DataFrame\n",
    "        obj_cols: list of str \n",
    "    Returns:\n",
    "        pd.DataFrame\n",
    "    '''\n",
    "    \n",
    "    cheat_count_is_cheater = df['is_cheater'].value_counts().index[0]\n",
    "    cheat_count_has_email = df['has_email'].value_counts().index[0]\n",
    "    \n",
    "    df['is_cheater'] = df['is_cheater'].fillna(cheat_count_is_cheater)\n",
    "    df['has_email'] = df['has_email'].fillna(cheat_count_has_email)\n",
    "    \n",
    "    df['country'] = df['country'].fillna('other')\n",
    "    df['network_name'] = df['network_name'].fillna('other')\n",
    "\n",
    "    encoder = OneHotEncoder(sparse=False, handle_unknown='ignore')\n",
    "    for column in obj_cols:\n",
    "        encoded_data = encoder.fit_transform(df[[column]])\n",
    "        encoded_df = pd.DataFrame(encoded_data, columns=encoder.get_feature_names_out([column]))\n",
    "        df = pd.concat([df, encoded_df], axis=1)\n",
    "    \n",
    "    df.drop(obj_cols, axis=1, inplace=True)\n",
    "        \n",
    "    return df\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/sklearn/preprocessing/_encoders.py:972: FutureWarning: `sparse` was renamed to `sparse_output` in version 1.2 and will be removed in 1.4. `sparse_output` is ignored unless you leave `sparse` to its default value.\n",
      "  warnings.warn(\n",
      "/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/sklearn/preprocessing/_encoders.py:972: FutureWarning: `sparse` was renamed to `sparse_output` in version 1.2 and will be removed in 1.4. `sparse_output` is ignored unless you leave `sparse` to its default value.\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "df = prepare_obj_data(df, obj_cols)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df.drop('target_game_currency', axis = 1)\n",
    "y = df['target_game_currency']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_stratified = (y > 0).astype(int)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.10, stratify=y_stratified, random_state=42)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Catboost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_importance(model, X):\n",
    "    '''\n",
    "    Function for plot regression coef\n",
    "    '''\n",
    "    try:\n",
    "        fold_importance_df = pd.DataFrame()\n",
    "        fold_importance_df[\"feature\"] = X.columns\n",
    "        fold_importance_df[\"importance\"] = model.feature_importances_\n",
    "        td = fold_importance_df.sort_values('importance', ascending=False)\n",
    "        sns.barplot(x=\"importance\", y=\"feature\", data=td.head(10), palette=\"Blues_d\")\n",
    "        plt.show()\n",
    "    except:\n",
    "        return []"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def SMAPE(y_true, y_pred):\n",
    "    \"\"\"\n",
    "    SMAPE score\n",
    "    \"\"\"\n",
    "    denominator = (np.abs(y_true) + np.abs(y_pred))\n",
    "    diff = np.where(denominator == 0, 0, 2 * np.abs(y_true - y_pred) / denominator)\n",
    "    return np.mean(diff)\n",
    "\n",
    "\n",
    "def custom_smape_score(estimator, X, y):\n",
    "    '''\n",
    "    SMAP for cross-validation\n",
    "    '''\n",
    "    y_pred = estimator.predict(X)\n",
    "    smape = SMAPE(y, y_pred)\n",
    "    return -smape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_model(model_obj, X_train, y_train, num_folds ,scoring,cross_val = False):\n",
    "    '''\n",
    "    Train model\n",
    "    '''\n",
    "    model = model_obj\n",
    "    print(model.fit(X_train, y_train, verbose=50, plot=True))\n",
    "    \n",
    "    if cross_val:\n",
    "        cv_scores = cross_val_score(model, X_train, y_train, cv = num_folds, scoring = scoring)\n",
    "        print(\"Cross_validation:\", cv_scores)\n",
    "    \n",
    "    td = plot_importance(model, X_train)\n",
    "    \n",
    "    return model, td"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "88905e926bf043af9236f43e89394e95",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "MetricVisualizer(layout=Layout(align_self='stretch', height='500px'))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0:\tlearn: 484.1682889\ttotal: 113ms\tremaining: 25.8s\n",
      "50:\tlearn: 433.8854273\ttotal: 5.76s\tremaining: 20.2s\n",
      "100:\tlearn: 428.9690557\ttotal: 10.3s\tremaining: 13.1s\n",
      "150:\tlearn: 428.4892535\ttotal: 12.8s\tremaining: 6.71s\n",
      "200:\tlearn: 427.9060742\ttotal: 17.6s\tremaining: 2.54s\n",
      "229:\tlearn: 427.5022304\ttotal: 21s\tremaining: 0us\n",
      "<catboost.core.CatBoostRegressor object at 0x2ea4d5290>\n",
      "0:\tlearn: 488.9098849\ttotal: 82.3ms\tremaining: 18.8s\n",
      "50:\tlearn: 437.3353180\ttotal: 4.84s\tremaining: 17s\n",
      "100:\tlearn: 431.6915648\ttotal: 9.32s\tremaining: 11.9s\n",
      "150:\tlearn: 430.5938317\ttotal: 14s\tremaining: 7.32s\n",
      "200:\tlearn: 430.1702938\ttotal: 18.6s\tremaining: 2.68s\n",
      "229:\tlearn: 430.0639143\ttotal: 21.3s\tremaining: 0us\n",
      "0:\tlearn: 486.9009092\ttotal: 93.9ms\tremaining: 21.5s\n",
      "50:\tlearn: 436.7346234\ttotal: 4.66s\tremaining: 16.3s\n",
      "100:\tlearn: 432.6951143\ttotal: 9.44s\tremaining: 12.1s\n",
      "150:\tlearn: 431.1393834\ttotal: 14.1s\tremaining: 7.4s\n",
      "200:\tlearn: 430.7188505\ttotal: 16.4s\tremaining: 2.36s\n",
      "229:\tlearn: 430.7100463\ttotal: 17.7s\tremaining: 0us\n",
      "0:\tlearn: 482.8489335\ttotal: 66.4ms\tremaining: 15.2s\n",
      "50:\tlearn: 432.8772796\ttotal: 3.98s\tremaining: 14s\n",
      "100:\tlearn: 428.3796872\ttotal: 8.54s\tremaining: 10.9s\n",
      "150:\tlearn: 427.7214164\ttotal: 13.5s\tremaining: 7.05s\n",
      "200:\tlearn: 427.2823732\ttotal: 20s\tremaining: 2.89s\n",
      "229:\tlearn: 427.0467304\ttotal: 23s\tremaining: 0us\n",
      "0:\tlearn: 489.9957884\ttotal: 66.5ms\tremaining: 15.2s\n",
      "50:\tlearn: 442.0240588\ttotal: 4.89s\tremaining: 17.2s\n",
      "100:\tlearn: 436.2372611\ttotal: 9.19s\tremaining: 11.7s\n",
      "150:\tlearn: 435.1489874\ttotal: 13.7s\tremaining: 7.14s\n",
      "200:\tlearn: 434.8422178\ttotal: 18.2s\tremaining: 2.62s\n",
      "229:\tlearn: 434.4679414\ttotal: 20.9s\tremaining: 0us\n",
      "0:\tlearn: 471.8314515\ttotal: 98.7ms\tremaining: 22.6s\n",
      "50:\tlearn: 421.8151064\ttotal: 4.57s\tremaining: 16s\n",
      "100:\tlearn: 419.4796734\ttotal: 7.98s\tremaining: 10.2s\n",
      "150:\tlearn: 417.6793637\ttotal: 10s\tremaining: 5.24s\n",
      "200:\tlearn: 417.5245545\ttotal: 12.7s\tremaining: 1.83s\n",
      "229:\tlearn: 417.5245545\ttotal: 15.6s\tremaining: 0us\n",
      "Cross_validation: [-417.19417448 -418.47560015 -434.83371395 -403.77859593 -478.60685987]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "params = {\n",
    "    'iterations': 230,\n",
    "    'depth': 3,\n",
    "    'loss_function': 'MAE',\n",
    "    'random_state': 42,\n",
    "    'l2_leaf_reg': 20,\n",
    "    'verbose': 50,\n",
    "    'early_stopping_rounds': 20,\n",
    "    'thread_count': -1\n",
    "}\n",
    "\n",
    "model, features = train_model(CatBoostRegressor(**params),\n",
    "                              X_train, \n",
    "                              y_train,\n",
    "                              5,\n",
    "                              scoring='neg_mean_absolute_error',\n",
    "                              cross_val=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_train = model.predict(X_train)\n",
    "y_pred_test = model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train SMAPE:  1.9775113990451205\n",
      "Test SMAPE:  1.9779116425657624\n",
      "Train RMSE:  8313.044680082789\n",
      "Test RMSE:  7197.929424031383\n",
      "Train MAE:  427.50223179291436\n",
      "Test MAE:  378.8722657599412\n"
     ]
    }
   ],
   "source": [
    "print('Train SMAPE: ', SMAPE(y_train, y_pred_train))\n",
    "print('Test SMAPE: ', SMAPE(y_test, y_pred_test))\n",
    "\n",
    "print('Train RMSE: ',  np.sqrt(mean_squared_error(y_train, y_pred_train)))\n",
    "print('Test RMSE: ',  np.sqrt(mean_squared_error(y_test, y_pred_test)))\n",
    "\n",
    "print('Train MAE: ',  mean_absolute_error(y_train, y_pred_train))\n",
    "print('Test MAE: ',  mean_absolute_error(y_test, y_pred_test))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "base",
   "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": "3.11.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}