XGBoost.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "import xgboost as xgb\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import confusion_matrix, accuracy_score, precision_score, recall_score, f1_score, roc_auc_score\n",
    "from sklearn.metrics import mean_absolute_error, mean_squared_error \n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "def total_score(y_test, pred = None, pred_proba= None):\n",
    "    confusion = confusion_matrix(y_test, pred)\n",
    "    accuracy = accuracy_score(y_test, pred)\n",
    "    precision = precision_score(y_test, pred)\n",
    "    recall = recall_score(y_test, pred)\n",
    "    f1 = f1_score(y_test, pred)\n",
    "    roc_auc = roc_auc_score(y_test,pred_proba)\n",
    "    print('오차 행렬')\n",
    "    print(confusion)\n",
    "    print('정확도:{0:.4f}, 정밀도:{1:.4f}, 재현율:{2:.4f}, F1:{3:,4f}, AUC:{4:.4f}'.format(accuracy,precision, recall, f1 , roc_auc))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "path = './data/data_for_model.csv' #./:작업하고 있는 파일을 나타내는 위치\n",
    "\n",
    "data = pd.read_csv(path)\n",
    "df=data\n",
    "del df[\"wind_speed\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = df.iloc[:,2:13]\n",
    "y = df.iloc[:,-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>#sk-container-id-2 {color: black;background-color: white;}#sk-container-id-2 pre{padding: 0;}#sk-container-id-2 div.sk-toggleable {background-color: white;}#sk-container-id-2 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-2 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-2 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-2 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-2 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-2 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-2 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-2 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-2 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-2 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-2 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-2 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-2 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-2 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-2 div.sk-item {position: relative;z-index: 1;}#sk-container-id-2 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-2 div.sk-item::before, #sk-container-id-2 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-2 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-2 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-2 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-2 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-2 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-2 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-2 div.sk-label-container {text-align: center;}#sk-container-id-2 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-2 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-2\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>XGBRegressor(base_score=0.5, booster=&#x27;gbtree&#x27;, callbacks=None,\n",
       "             colsample_bylevel=1, colsample_bynode=1, colsample_bytree=1,\n",
       "             early_stopping_rounds=None, enable_categorical=False,\n",
       "             eval_metric=None, feature_types=None, gamma=0, gpu_id=-1,\n",
       "             grow_policy=&#x27;depthwise&#x27;, importance_type=None,\n",
       "             interaction_constraints=&#x27;&#x27;, learning_rate=0.300000012, max_bin=256,\n",
       "             max_cat_threshold=64, max_cat_to_onehot=4, max_delta_step=0,\n",
       "             max_depth=6, max_leaves=0, min_child_weight=1, missing=nan,\n",
       "             monotone_constraints=&#x27;()&#x27;, n_estimators=100, n_jobs=0,\n",
       "             num_parallel_tree=1, predictor=&#x27;auto&#x27;, random_state=0, ...)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-2\" type=\"checkbox\" checked><label for=\"sk-estimator-id-2\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">XGBRegressor</label><div class=\"sk-toggleable__content\"><pre>XGBRegressor(base_score=0.5, booster=&#x27;gbtree&#x27;, callbacks=None,\n",
       "             colsample_bylevel=1, colsample_bynode=1, colsample_bytree=1,\n",
       "             early_stopping_rounds=None, enable_categorical=False,\n",
       "             eval_metric=None, feature_types=None, gamma=0, gpu_id=-1,\n",
       "             grow_policy=&#x27;depthwise&#x27;, importance_type=None,\n",
       "             interaction_constraints=&#x27;&#x27;, learning_rate=0.300000012, max_bin=256,\n",
       "             max_cat_threshold=64, max_cat_to_onehot=4, max_delta_step=0,\n",
       "             max_depth=6, max_leaves=0, min_child_weight=1, missing=nan,\n",
       "             monotone_constraints=&#x27;()&#x27;, n_estimators=100, n_jobs=0,\n",
       "             num_parallel_tree=1, predictor=&#x27;auto&#x27;, random_state=0, ...)</pre></div></div></div></div></div>"
      ],
      "text/plain": [
       "XGBRegressor(base_score=0.5, booster='gbtree', callbacks=None,\n",
       "             colsample_bylevel=1, colsample_bynode=1, colsample_bytree=1,\n",
       "             early_stopping_rounds=None, enable_categorical=False,\n",
       "             eval_metric=None, feature_types=None, gamma=0, gpu_id=-1,\n",
       "             grow_policy='depthwise', importance_type=None,\n",
       "             interaction_constraints='', learning_rate=0.300000012, max_bin=256,\n",
       "             max_cat_threshold=64, max_cat_to_onehot=4, max_delta_step=0,\n",
       "             max_depth=6, max_leaves=0, min_child_weight=1, missing=nan,\n",
       "             monotone_constraints='()', n_estimators=100, n_jobs=0,\n",
       "             num_parallel_tree=1, predictor='auto', random_state=0, ...)"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = xgb.XGBRegressor()\n",
    "\n",
    "model.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = model.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.5778713199856953\n",
      "0.7418284352903624\n"
     ]
    }
   ],
   "source": [
    "mae = mean_absolute_error(y_test, predictions)\n",
    "mse = mean_squared_error(y_test, predictions)\n",
    "\n",
    "print(mae)\n",
    "print(mse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot: title={'center': 'Feature importance'}, xlabel='F score', ylabel='Features'>"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "xgb.plot_importance(model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9975512275083803\n",
      "0.9964379953800687\n"
     ]
    }
   ],
   "source": [
    "print(model.score(X_train, y_train))\n",
    "print(model.score(X_test, y_test))"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Gamma에 따른 튜닝"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "model1 = xgb.XGBRegressor(learning_rate = 0.3, n_estimators = 100, min_child_weight = 100,\n",
    "                          gamma = 10, max_depth = 6, colsample_bytree = 0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9951735333370135\n",
      "0.9940302539507583\n",
      "0.7898843082261481\n",
      "1.2432682838225206\n"
     ]
    }
   ],
   "source": [
    "model1.fit(X_train, y_train)\n",
    "predictions = model1.predict(X_test)\n",
    "\n",
    "mae = mean_absolute_error(y_test, predictions)\n",
    "mse = mean_squared_error(y_test, predictions)\n",
    "\n",
    "print(model1.score(X_train, y_train))\n",
    "print(model1.score(X_test, y_test))\n",
    "print(mae)\n",
    "print(mse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9949665970883295\n",
      "0.9938671157120267\n",
      "0.8079712102739365\n",
      "1.2772436985923719\n"
     ]
    }
   ],
   "source": [
    "model = xgb.XGBRegressor(learning_rate = 0.3, n_estimators = 100, min_child_weight = 100,\n",
    "                          gamma = 20, max_depth = 6, colsample_bytree = 0.5)\n",
    "model.fit(X_train,y_train)\n",
    "predictions = model.predict(X_test)\n",
    "mae = mean_absolute_error(y_test, predictions)\n",
    "mse = mean_squared_error(y_test, predictions)\n",
    "\n",
    "print(model.score(X_train, y_train))\n",
    "print(model.score(X_test, y_test))\n",
    "print(mae)\n",
    "print(mse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9933856233914483\n",
      "0.9923066967485363\n",
      "0.9017873715714786\n",
      "1.6022188969978646\n"
     ]
    }
   ],
   "source": [
    "model = xgb.XGBRegressor(learning_rate = 0.3, n_estimators = 100, min_child_weight = 100,\n",
    "                          gamma = 100, max_depth = 6, colsample_bytree = 0.5)\n",
    "model.fit(X_train,y_train)\n",
    "predictions = model.predict(X_test)\n",
    "mae = mean_absolute_error(y_test, predictions)\n",
    "mse = mean_squared_error(y_test, predictions)\n",
    "\n",
    "print(model.score(X_train, y_train))\n",
    "print(model.score(X_test, y_test))\n",
    "print(mae)\n",
    "print(mse)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "n_estimator에 따른 튜닝"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9973770089048657\n",
      "0.9961867760509012\n",
      "0.6294749949416738\n",
      "0.7941477503266916\n"
     ]
    }
   ],
   "source": [
    "model = xgb.XGBRegressor(learning_rate = 0.3, n_estimators = 1000, min_child_weight = 100,\n",
    "                          gamma = 10, max_depth = 6, colsample_bytree = 0.5)\n",
    "model.fit(X_train,y_train)\n",
    "predictions = model.predict(X_test)\n",
    "mae = mean_absolute_error(y_test, predictions)\n",
    "mse = mean_squared_error(y_test, predictions)\n",
    "\n",
    "print(model.score(X_train, y_train))\n",
    "print(model.score(X_test, y_test))\n",
    "print(mae)\n",
    "print(mse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9973770089232658\n",
      "0.9961867760501492\n",
      "0.6294749947934313\n",
      "0.7941477504833196\n"
     ]
    }
   ],
   "source": [
    "model = xgb.XGBRegressor(learning_rate = 0.3, n_estimators = 10000, min_child_weight = 100,\n",
    "                          gamma = 10, max_depth = 6, colsample_bytree = 0.5)\n",
    "model.fit(X_train,y_train)\n",
    "predictions = model.predict(X_test)\n",
    "mae = mean_absolute_error(y_test, predictions)\n",
    "mse = mean_squared_error(y_test, predictions)\n",
    "\n",
    "print(model.score(X_train, y_train))\n",
    "print(model.score(X_test, y_test))\n",
    "print(mae)\n",
    "print(mse)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "max_depth에 따른 튜닝"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9973494607038506\n",
      "0.9949787251087613\n",
      "0.7560227415428401\n",
      "1.0457382550509586\n"
     ]
    }
   ],
   "source": [
    "model = xgb.XGBRegressor(learning_rate = 0.3, n_estimators = 100, min_child_weight = 100,\n",
    "                          gamma = 10, max_depth = 20, colsample_bytree = 0.5)\n",
    "model.fit(X_train,y_train)\n",
    "predictions = model.predict(X_test)\n",
    "mae = mean_absolute_error(y_test, predictions)\n",
    "mse = mean_squared_error(y_test, predictions)\n",
    "\n",
    "print(model.score(X_train, y_train))\n",
    "print(model.score(X_test, y_test))\n",
    "print(mae)\n",
    "print(mse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9700044629496475\n",
      "0.9696871392774413\n",
      "1.8553745131823454\n",
      "6.313001929633738\n"
     ]
    }
   ],
   "source": [
    "model = xgb.XGBRegressor(learning_rate = 0.3, n_estimators = 100, min_child_weight = 100,\n",
    "                          gamma = 10, max_depth = 1, colsample_bytree = 0.5)\n",
    "model.fit(X_train,y_train)\n",
    "predictions = model.predict(X_test)\n",
    "mae = mean_absolute_error(y_test, predictions)\n",
    "mse = mean_squared_error(y_test, predictions)\n",
    "\n",
    "print(model.score(X_train, y_train))\n",
    "print(model.score(X_test, y_test))\n",
    "print(mae)\n",
    "print(mse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9899777926404146\n",
      "0.9893696031664645\n",
      "1.0301318817127214\n",
      "2.213902420398729\n"
     ]
    }
   ],
   "source": [
    "model = xgb.XGBRegressor(learning_rate = 0.3, n_estimators = 100, min_child_weight = 100,\n",
    "                          gamma = 10, max_depth = 3, colsample_bytree = 0.5)\n",
    "model.fit(X_train,y_train)\n",
    "predictions = model.predict(X_test)\n",
    "mae = mean_absolute_error(y_test, predictions)\n",
    "mse = mean_squared_error(y_test, predictions)\n",
    "\n",
    "print(model.score(X_train, y_train))\n",
    "print(model.score(X_test, y_test))\n",
    "print(mae)\n",
    "print(mse)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "min_child_weight에 따른 튜닝"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9958243651043416\n",
      "0.9944828517479151\n",
      "0.765404935464321\n",
      "1.1490095864019587\n"
     ]
    }
   ],
   "source": [
    "model = xgb.XGBRegressor(learning_rate = 0.3, n_estimators = 100, min_child_weight = 10,\n",
    "                          gamma = 10, max_depth = 6, colsample_bytree = 0.5)\n",
    "model.fit(X_train,y_train)\n",
    "predictions = model.predict(X_test)\n",
    "mae = mean_absolute_error(y_test, predictions)\n",
    "mse = mean_squared_error(y_test, predictions)\n",
    "\n",
    "print(model.score(X_train, y_train))\n",
    "print(model.score(X_test, y_test))\n",
    "print(mae)\n",
    "print(mse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9919390801836206\n",
      "0.9912454285491179\n",
      "0.9495098514972758\n",
      "1.8232402071311347\n"
     ]
    }
   ],
   "source": [
    "model = xgb.XGBRegressor(learning_rate = 0.3, n_estimators = 100, min_child_weight = 1000,\n",
    "                          gamma = 10, max_depth = 6, colsample_bytree = 0.5)\n",
    "model.fit(X_train,y_train)\n",
    "predictions = model.predict(X_test)\n",
    "mae = mean_absolute_error(y_test, predictions)\n",
    "mse = mean_squared_error(y_test, predictions)\n",
    "\n",
    "print(model.score(X_train, y_train))\n",
    "print(model.score(X_test, y_test))\n",
    "print(mae)\n",
    "print(mse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9935733989514763\n",
      "0.9926892813805077\n",
      "0.867764225994578\n",
      "1.52254124657781\n"
     ]
    }
   ],
   "source": [
    "model = xgb.XGBRegressor(learning_rate = 0.3, n_estimators = 100, min_child_weight = 500,\n",
    "                          gamma = 10, max_depth = 6, colsample_bytree = 0.5)\n",
    "model.fit(X_train,y_train)\n",
    "predictions = model.predict(X_test)\n",
    "mae = mean_absolute_error(y_test, predictions)\n",
    "mse = mean_squared_error(y_test, predictions)\n",
    "\n",
    "print(model.score(X_train, y_train))\n",
    "print(model.score(X_test, y_test))\n",
    "print(mae)\n",
    "print(mse)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "colsample_bytree에 따른 튜닝"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9965926313214072\n",
      "0.9956406431401006\n",
      "0.6328529135281337\n",
      "0.9078862110835426\n"
     ]
    }
   ],
   "source": [
    "model = xgb.XGBRegressor(learning_rate = 0.3, n_estimators = 100, min_child_weight = 100,\n",
    "                          gamma = 10, max_depth = 6, colsample_bytree = 1)\n",
    "model.fit(X_train,y_train)\n",
    "predictions = model.predict(X_test)\n",
    "mae = mean_absolute_error(y_test, predictions)\n",
    "mse = mean_squared_error(y_test, predictions)\n",
    "\n",
    "print(model.score(X_train, y_train))\n",
    "print(model.score(X_test, y_test))\n",
    "print(mae)\n",
    "print(mse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9926452204448667\n",
      "0.9913767910773326\n",
      "0.9502447025341952\n",
      "1.795882449587536\n"
     ]
    }
   ],
   "source": [
    "model = xgb.XGBRegressor(learning_rate = 0.3, n_estimators = 100, min_child_weight = 100,\n",
    "                          gamma = 10, max_depth = 6, colsample_bytree = 0.25)\n",
    "model.fit(X_train,y_train)\n",
    "predictions = model.predict(X_test)\n",
    "mae = mean_absolute_error(y_test, predictions)\n",
    "mse = mean_squared_error(y_test, predictions)\n",
    "\n",
    "print(model.score(X_train, y_train))\n",
    "print(model.score(X_test, y_test))\n",
    "print(mae)\n",
    "print(mse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9951735333370135\n",
      "0.9940302539507583\n",
      "0.7898843082261481\n",
      "1.2432682838225206\n"
     ]
    }
   ],
   "source": [
    "model = xgb.XGBRegressor(learning_rate = 0.3, n_estimators = 100, min_child_weight = 100,\n",
    "                          gamma = 10, max_depth = 6, colsample_bytree = 0.5)\n",
    "model.fit(X_train,y_train)\n",
    "predictions = model.predict(X_test)\n",
    "mae = mean_absolute_error(y_test, predictions)\n",
    "mse = mean_squared_error(y_test, predictions)\n",
    "\n",
    "print(model.score(X_train, y_train))\n",
    "print(model.score(X_test, y_test))\n",
    "print(mae)\n",
    "print(mse)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "learning_rate에 따른 튜닝"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9944050395879094\n",
      "0.992663350574736\n",
      "0.8751683229380235\n",
      "1.5279416351578445\n"
     ]
    }
   ],
   "source": [
    "model = xgb.XGBRegressor(learning_rate = 0.5, n_estimators = 100, min_child_weight = 100,\n",
    "                          gamma = 10, max_depth = 6, colsample_bytree = 0.5)\n",
    "model.fit(X_train,y_train)\n",
    "predictions = model.predict(X_test)\n",
    "mae = mean_absolute_error(y_test, predictions)\n",
    "mse = mean_squared_error(y_test, predictions)\n",
    "\n",
    "print(model.score(X_train, y_train))\n",
    "print(model.score(X_test, y_test))\n",
    "print(mae)\n",
    "print(mse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9919138710418047\n",
      "0.9886698114881267\n",
      "1.091815773590372\n",
      "2.3596420870082846\n"
     ]
    }
   ],
   "source": [
    "model = xgb.XGBRegressor(learning_rate = 0.75, n_estimators = 100, min_child_weight = 100,\n",
    "                          gamma = 10, max_depth = 6, colsample_bytree = 0.5)\n",
    "model.fit(X_train,y_train)\n",
    "predictions = model.predict(X_test)\n",
    "mae = mean_absolute_error(y_test, predictions)\n",
    "mse = mean_squared_error(y_test, predictions)\n",
    "\n",
    "print(model.score(X_train, y_train))\n",
    "print(model.score(X_test, y_test))\n",
    "print(mae)\n",
    "print(mse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "#eta: learning_rate >> 0~1의 범위, 클수록 과적합\n",
    "#num_boost_aroung(n_estimators): 학습에 사용뢰 weak learner의 수\n",
    "#min_child_weight: leaf node에 포함될 최고 관측치의 수 >> 0~무한대, 작을수록 과적합\n",
    "#gamma: leaf node의 추가분할을 결정할 최소손실 감소값 >> 0~무한대, 클수록 과적합\n",
    "#max_depth: 트리의 최대 깊이 설정 >> 0~무한대, 과적합에 가장 민감하게 작용함\n",
    "#sub_sample: 학습 시 데이터 샘플링 비율을 지정 >> 0~1이지만 보통 0.5~1 사용, 미리 split을 해놓으면 의미는 없는듯?\n",
    "#colsample_bytree: 트리 생성에 필요한 feature의 샘플링에 사용 >> 0~1, feature가 많은 때 과적합 조절에 이용"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "venv",
   "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": "3.8.10"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "69089f81ed2b007dd37b053b0ecff544931d6f38c1b933baa1b3faccb026f3e3"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}