5_clearml_train.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from math import sqrt\n",
    "\n",
    "from clearml import Task\n",
    "import argparse\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"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ClearML Task: created new task id=4591cc2341c74217a5ccf2943ee18794\n",
      "2023-12-29 18:49:33,840 - clearml.Task - INFO - Storing jupyter notebook directly as code\n",
      "ClearML results page: https://app.clear.ml/projects/1db2548eb69c4c26883ef321cdc663bf/experiments/4591cc2341c74217a5ccf2943ee18794/output/log\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ClearML Monitor: GPU monitoring failed getting GPU reading, switching off GPU monitoring\n",
      "ClearML Monitor: Could not detect iteration reporting, falling back to iterations as seconds-from-start\n"
     ]
    }
   ],
   "source": [
    "task = Task.init(project_name=\"Dataset_g5\", task_name=\"Catboost_train\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "parser = argparse.ArgumentParser()\n",
    "parser.add_argument(\"-iterations\", default=30, help=\"iterations boosting\")\n",
    "parser.add_argument(\"-depth\", default=3, help=\"depth tree\")\n",
    "parser.add_argument(\"-loss_function\", default=\"MAE\", help=\"loss function\")\n",
    "\n",
    "args = parser.parse_known_args()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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(\n",
    "            encoded_data, columns=encoder.get_feature_names_out([column])\n",
    "        )\n",
    "        df = pd.concat([df, encoded_df], axis=1)\n",
    "\n",
    "    df.drop(obj_cols, axis=1, inplace=True)\n",
    "\n",
    "    return df"
   ]
  },
  {
   "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 []\n",
    "\n",
    "\n",
    "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\n",
    "\n",
    "\n",
    "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(\n",
    "            model, X_train, y_train, cv=num_folds, scoring=scoring\n",
    "        )\n",
    "        print(\"Cross_validation:\", cv_scores)\n",
    "\n",
    "    td = plot_importance(model, X_train)\n",
    "\n",
    "    return model, td"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv(\"dataset.csv\", index_col=0)\n",
    "\n",
    "df.drop(\n",
    "    [\"date_install\", \"first_command_time\", \"birthday\", \"sex\", \"time_confirm_email\"],\n",
    "    axis=1,\n",
    "    inplace=True,\n",
    ")\n",
    "int_cols = df.select_dtypes(include=[\"int\", \"float\"]).columns\n",
    "int_cols = int_cols.drop(\"target_game_currency\")\n",
    "obj_cols = [\"country\", \"network_name\"]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/kurmangazykarabekov/anaconda3/lib/python3.11/site-packages/sklearn/preprocessing/_encoders.py:972: FutureWarning:\n",
      "\n",
      "`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",
      "\n",
      "/Users/kurmangazykarabekov/anaconda3/lib/python3.11/site-packages/sklearn/preprocessing/_encoders.py:972: FutureWarning:\n",
      "\n",
      "`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",
      "\n"
     ]
    }
   ],
   "source": [
    "df = prepare_obj_data(df, obj_cols)\n",
    "X = df.drop(\"target_game_currency\", axis=1)\n",
    "y = df[\"target_game_currency\"]\n",
    "\n",
    "y_stratified = (y > 0).astype(int)\n",
    "X_train, X_test, y_train, y_test = train_test_split(\n",
    "    X, y, test_size=0.10, stratify=y_stratified, random_state=42\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "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",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'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}"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "task.connect(params)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "16a3b3097cb8453d80823cd2b7695774",
       "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: 129ms\tremaining: 29.5s\n",
      "50:\tlearn: 433.8854273\ttotal: 2.35s\tremaining: 8.25s\n",
      "100:\tlearn: 428.9690557\ttotal: 4.45s\tremaining: 5.69s\n",
      "150:\tlearn: 428.4892535\ttotal: 6.67s\tremaining: 3.49s\n",
      "200:\tlearn: 427.9060742\ttotal: 8.79s\tremaining: 1.27s\n",
      "229:\tlearn: 427.5022304\ttotal: 10.1s\tremaining: 0us\n",
      "<catboost.core.CatBoostRegressor object at 0x17b402d50>\n",
      "0:\tlearn: 488.9098849\ttotal: 65ms\tremaining: 14.9s\n",
      "50:\tlearn: 437.3353180\ttotal: 2.22s\tremaining: 7.79s\n",
      "100:\tlearn: 431.6915648\ttotal: 3.89s\tremaining: 4.97s\n",
      "150:\tlearn: 430.5938317\ttotal: 6.27s\tremaining: 3.28s\n",
      "200:\tlearn: 430.1702938\ttotal: 8.27s\tremaining: 1.19s\n",
      "229:\tlearn: 430.0639143\ttotal: 9.24s\tremaining: 0us\n",
      "0:\tlearn: 486.9009092\ttotal: 92.3ms\tremaining: 21.1s\n",
      "50:\tlearn: 436.7346234\ttotal: 1.95s\tremaining: 6.85s\n",
      "100:\tlearn: 432.6951143\ttotal: 3.65s\tremaining: 4.66s\n",
      "150:\tlearn: 431.1393834\ttotal: 5.3s\tremaining: 2.77s\n",
      "200:\tlearn: 430.7188505\ttotal: 7.01s\tremaining: 1.01s\n",
      "229:\tlearn: 430.7100463\ttotal: 8.07s\tremaining: 0us\n",
      "0:\tlearn: 482.8489335\ttotal: 43.5ms\tremaining: 9.96s\n",
      "50:\tlearn: 432.8772796\ttotal: 1.82s\tremaining: 6.4s\n",
      "100:\tlearn: 428.3796872\ttotal: 3.5s\tremaining: 4.47s\n",
      "150:\tlearn: 427.7214164\ttotal: 5.17s\tremaining: 2.7s\n",
      "200:\tlearn: 427.2823732\ttotal: 6.82s\tremaining: 984ms\n",
      "229:\tlearn: 427.0467304\ttotal: 7.77s\tremaining: 0us\n",
      "0:\tlearn: 489.9957884\ttotal: 42ms\tremaining: 9.61s\n",
      "50:\tlearn: 442.0240588\ttotal: 1.73s\tremaining: 6.07s\n",
      "100:\tlearn: 436.2372611\ttotal: 3.38s\tremaining: 4.32s\n",
      "150:\tlearn: 435.1489874\ttotal: 5.06s\tremaining: 2.65s\n",
      "200:\tlearn: 434.8422178\ttotal: 6.73s\tremaining: 971ms\n",
      "229:\tlearn: 434.4679414\ttotal: 7.69s\tremaining: 0us\n",
      "0:\tlearn: 471.8314515\ttotal: 65.4ms\tremaining: 15s\n",
      "50:\tlearn: 421.8151064\ttotal: 1.75s\tremaining: 6.14s\n",
      "100:\tlearn: 419.4796734\ttotal: 3.4s\tremaining: 4.34s\n",
      "150:\tlearn: 417.6793637\ttotal: 5.04s\tremaining: 2.64s\n",
      "200:\tlearn: 417.5245545\ttotal: 6.69s\tremaining: 965ms\n",
      "229:\tlearn: 417.5245545\ttotal: 7.64s\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": [
    "\n",
    "\n",
    "model, features = train_model(\n",
    "    CatBoostRegressor(**params),\n",
    "    X_train,\n",
    "    y_train,\n",
    "    5,\n",
    "    scoring=\"neg_mean_absolute_error\",\n",
    "    cross_val=True,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.11.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}