ref_LinearRegression_Ridge_Rasso.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "936e2592",
   "metadata": {},
   "outputs": [],
   "source": [
    "# data_3_lightGBM을 활용해 선형회귀, 릿지회귀, 라쏘회귀 수행\n",
    "# 수행결과, 트리기반 모델에 비해 낮은 평가지표를 보임\n",
    "# 모델 자체의 성능이 아니라 전처리된 데이터 셋이 해당 모델이 적합하지 않은것으로 판단\n",
    "# 선형 회귀의 경우, 변수간의 독립성이 필요하지만, 트리기반 모델을 사용하면서 이를 고려치 않고 피쳐 생성\n",
    "# 실제로, 변수 선택이 가능한 라쏘의 경우, 선형회귀나 릿지에 비해 상당히 평가지표가 개선된 것을 확인 가능\n",
    "# PCA 분석으로 주성분을 결정해 동일 모델로 재수행한다면 더 좋은 결과를 예상"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "985157be",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.linear_model import LinearRegression, Ridge, Lasso\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.metrics import mean_squared_error as mse"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "467f7686",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 8029111 entries, 827196 to 8856306\n",
      "Data columns (total 29 columns):\n",
      " #   Column                      Dtype  \n",
      "---  ------                      -----  \n",
      " 0   month_id                    int8   \n",
      " 1   shop_id                     int8   \n",
      " 2   item_id                     int16  \n",
      " 3   cnt_month                   float32\n",
      " 4   city_id                     int8   \n",
      " 5   cat_id                      int8   \n",
      " 6   main_cat                    int8   \n",
      " 7   sub_cat                     int8   \n",
      " 8   year                        int16  \n",
      " 9   month                       int8   \n",
      " 10  sales_month                 int8   \n",
      " 11  cnt_month_1                 float32\n",
      " 12  cnt_month_2                 float32\n",
      " 13  cnt_month_3                 float32\n",
      " 14  item_id_mean_cnt_1          float32\n",
      " 15  item_id_mean_cnt_2          float32\n",
      " 16  item_id_mean_cnt_3          float32\n",
      " 17  sub_cat_mean_cnt_1          float32\n",
      " 18  sub_cat_mean_cnt_2          float32\n",
      " 19  sub_cat_mean_cnt_3          float32\n",
      " 20  price_mean_mean_cnt_1       float32\n",
      " 21  price_mean_mean_cnt_2       float32\n",
      " 22  price_mean_mean_cnt_3       float32\n",
      " 23  item_id/city_id_mean_cnt_1  float32\n",
      " 24  item_id/city_id_mean_cnt_2  float32\n",
      " 25  item_id/city_id_mean_cnt_3  float32\n",
      " 26  shop_id/cat_id_mean_cnt_1   float32\n",
      " 27  shop_id/cat_id_mean_cnt_2   float32\n",
      " 28  shop_id/cat_id_mean_cnt_3   float32\n",
      "dtypes: float32(19), int16(2), int8(8)\n",
      "memory usage: 735.1 MB\n"
     ]
    }
   ],
   "source": [
    "data = pd.read_pickle('./data/data_3.pkl')\n",
    "data.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "f275b6b8",
   "metadata": {},
   "outputs": [],
   "source": [
    "data[['cnt_month','cnt_month_1','cnt_month_2','cnt_month_3']]=data[['cnt_month','cnt_month_1','cnt_month_2','cnt_month_3']].clip(0,20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "15fc0433",
   "metadata": {},
   "outputs": [],
   "source": [
    "# train/validation/test data split\n",
    "X_train = data[data['month_id'] < 33].drop(['cnt_month'], axis=1)\n",
    "y_train = data[data['month_id'] < 33]['cnt_month']\n",
    "\n",
    "X_valid = data[data['month_id'] == 33].drop(['cnt_month'], axis=1)\n",
    "y_valid = data[data['month_id'] == 33]['cnt_month']\n",
    "\n",
    "X_test = data[data['month_id'] == 34].drop(['cnt_month'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "9c1e4a95",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE Value For Linear Regression:  1.3752245\n"
     ]
    }
   ],
   "source": [
    "# 선형회귀 모델을 초기화\n",
    "lModel = LinearRegression()\n",
    "\n",
    "# 모델을 학습시킨다.\n",
    "y_train_log = np.log1p(y_train) #log +1 \n",
    "y_valid_log = np.log1p(y_valid) #log +1 \n",
    "\n",
    "lModel.fit(X_train, y_train_log)\n",
    "\n",
    "# 예측하고 정확도를 평가한다.\n",
    "preds = lModel.predict(X_valid)\n",
    "print (\"RMSE Value For Linear Regression: \",\n",
    "       np.sqrt(mse(np.exp(y_valid_log),np.exp(preds))))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "b072f100",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'alpha': 0.01, 'max_iter': 3000}\n",
      "RMSE Value For Ridge Regression:  1.2330824\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>mean_fit_time</th>\n",
       "      <th>std_fit_time</th>\n",
       "      <th>mean_score_time</th>\n",
       "      <th>std_score_time</th>\n",
       "      <th>param_alpha</th>\n",
       "      <th>param_max_iter</th>\n",
       "      <th>params</th>\n",
       "      <th>split0_test_score</th>\n",
       "      <th>split1_test_score</th>\n",
       "      <th>split2_test_score</th>\n",
       "      <th>split3_test_score</th>\n",
       "      <th>split4_test_score</th>\n",
       "      <th>mean_test_score</th>\n",
       "      <th>std_test_score</th>\n",
       "      <th>rank_test_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2.666357</td>\n",
       "      <td>0.115270</td>\n",
       "      <td>0.157903</td>\n",
       "      <td>0.016177</td>\n",
       "      <td>0.01</td>\n",
       "      <td>3000</td>\n",
       "      <td>{'alpha': 0.01, 'max_iter': 3000}</td>\n",
       "      <td>-0.319907</td>\n",
       "      <td>-0.333720</td>\n",
       "      <td>-0.308792</td>\n",
       "      <td>-0.331414</td>\n",
       "      <td>-0.307177</td>\n",
       "      <td>-0.320202</td>\n",
       "      <td>0.011031</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2.766894</td>\n",
       "      <td>0.186152</td>\n",
       "      <td>0.153905</td>\n",
       "      <td>0.007123</td>\n",
       "      <td>0.1</td>\n",
       "      <td>3000</td>\n",
       "      <td>{'alpha': 0.1, 'max_iter': 3000}</td>\n",
       "      <td>-0.320821</td>\n",
       "      <td>-0.334402</td>\n",
       "      <td>-0.309202</td>\n",
       "      <td>-0.332085</td>\n",
       "      <td>-0.306891</td>\n",
       "      <td>-0.320680</td>\n",
       "      <td>0.011316</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2.406317</td>\n",
       "      <td>0.194428</td>\n",
       "      <td>0.147909</td>\n",
       "      <td>0.005212</td>\n",
       "      <td>1</td>\n",
       "      <td>3000</td>\n",
       "      <td>{'alpha': 1, 'max_iter': 3000}</td>\n",
       "      <td>-0.331717</td>\n",
       "      <td>-0.346895</td>\n",
       "      <td>-0.318800</td>\n",
       "      <td>-0.344515</td>\n",
       "      <td>-0.313837</td>\n",
       "      <td>-0.331153</td>\n",
       "      <td>0.013260</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2.310176</td>\n",
       "      <td>0.089941</td>\n",
       "      <td>0.151307</td>\n",
       "      <td>0.002798</td>\n",
       "      <td>2</td>\n",
       "      <td>3000</td>\n",
       "      <td>{'alpha': 2, 'max_iter': 3000}</td>\n",
       "      <td>-0.340535</td>\n",
       "      <td>-0.357129</td>\n",
       "      <td>-0.327796</td>\n",
       "      <td>-0.354344</td>\n",
       "      <td>-0.321016</td>\n",
       "      <td>-0.340164</td>\n",
       "      <td>0.014203</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2.741910</td>\n",
       "      <td>0.241460</td>\n",
       "      <td>0.168497</td>\n",
       "      <td>0.019426</td>\n",
       "      <td>3</td>\n",
       "      <td>3000</td>\n",
       "      <td>{'alpha': 3, 'max_iter': 3000}</td>\n",
       "      <td>-0.346602</td>\n",
       "      <td>-0.364047</td>\n",
       "      <td>-0.334121</td>\n",
       "      <td>-0.360993</td>\n",
       "      <td>-0.326270</td>\n",
       "      <td>-0.346407</td>\n",
       "      <td>0.014700</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   mean_fit_time  std_fit_time  mean_score_time  std_score_time param_alpha  \\\n",
       "0       2.666357      0.115270         0.157903        0.016177        0.01   \n",
       "1       2.766894      0.186152         0.153905        0.007123         0.1   \n",
       "2       2.406317      0.194428         0.147909        0.005212           1   \n",
       "3       2.310176      0.089941         0.151307        0.002798           2   \n",
       "4       2.741910      0.241460         0.168497        0.019426           3   \n",
       "\n",
       "  param_max_iter                             params  split0_test_score  \\\n",
       "0           3000  {'alpha': 0.01, 'max_iter': 3000}          -0.319907   \n",
       "1           3000   {'alpha': 0.1, 'max_iter': 3000}          -0.320821   \n",
       "2           3000     {'alpha': 1, 'max_iter': 3000}          -0.331717   \n",
       "3           3000     {'alpha': 2, 'max_iter': 3000}          -0.340535   \n",
       "4           3000     {'alpha': 3, 'max_iter': 3000}          -0.346602   \n",
       "\n",
       "   split1_test_score  split2_test_score  split3_test_score  split4_test_score  \\\n",
       "0          -0.333720          -0.308792          -0.331414          -0.307177   \n",
       "1          -0.334402          -0.309202          -0.332085          -0.306891   \n",
       "2          -0.346895          -0.318800          -0.344515          -0.313837   \n",
       "3          -0.357129          -0.327796          -0.354344          -0.321016   \n",
       "4          -0.364047          -0.334121          -0.360993          -0.326270   \n",
       "\n",
       "   mean_test_score  std_test_score  rank_test_score  \n",
       "0        -0.320202        0.011031                1  \n",
       "1        -0.320680        0.011316                2  \n",
       "2        -0.331153        0.013260                3  \n",
       "3        -0.340164        0.014203                4  \n",
       "4        -0.346407        0.014700                5  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ridge_m_ = Ridge(normalize=True)\n",
    "ridge_params_ = { 'max_iter':[3000],'alpha':[0.01, 0.1, 1, 2, 3, 4, 10, 30,100,200,300,400,800,900,1000]}\n",
    "grid_ridge_m = GridSearchCV( ridge_m_,\n",
    "                          ridge_params_,\n",
    "                          scoring ='neg_root_mean_squared_error',\n",
    "                          cv=5)\n",
    "\n",
    "y_train_log = np.log1p(y_train)\n",
    "y_valid_log = np.log1p(y_valid) #log +1 \n",
    "\n",
    "grid_ridge_m.fit( X_train, y_train_log )\n",
    "preds = grid_ridge_m.predict(X_valid)\n",
    "print (grid_ridge_m.best_params_)\n",
    "print (\"RMSE Value For Ridge Regression: \", np.sqrt(mse(np.exp(y_valid_log),np.exp(preds))))\n",
    "\n",
    "df = pd.DataFrame(grid_ridge_m.cv_results_)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "e7c639aa",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'alpha': 10.0, 'max_iter': 3000}\n",
      "RMSE Value For Lasso Regression:  1.084341\n"
     ]
    }
   ],
   "source": [
    "lasso_m_ = Lasso(normalize=True)\n",
    "\n",
    "alpha  = 1/np.array([0.1, 1, 2, 3, 4, 10, 30,100,200,300,400,800,900,1000])\n",
    "lasso_params_ = { 'max_iter':[3000],'alpha':alpha}\n",
    "\n",
    "grid_lasso_m = GridSearchCV( lasso_m_,lasso_params_,scoring = 'neg_root_mean_squared_error',cv=5)\n",
    "y_train_log = np.log1p(y_train)\n",
    "y_valid_log = np.log1p(y_valid) #log +1 \n",
    "\n",
    "grid_lasso_m.fit( X_train , y_train_log )\n",
    "preds = grid_lasso_m.predict(X_valid)\n",
    "print (grid_lasso_m.best_params_)\n",
    "print (\"RMSE Value For Lasso Regression: \", np.sqrt(mse(np.exp(y_valid_log),np.exp(preds))))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}