Python87-LightGBM调参实例.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 实例1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "from sklearn.metrics import roc_auc_score\n",
    "import lightgbm as lgb\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# sklearn数据预处理\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.model_selection import PredefinedSplit\n",
    "from sklearn.model_selection import GridSearchCV,ParameterGrid\n",
    "from sklearn.metrics import roc_curve,auc,accuracy_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 参数设置\n",
    "params = {\n",
    "    'task': 'train',\n",
    "    'boosting_type': 'gbdt',\n",
    "    'objective': 'binary',\n",
    "    'metric': {'binary_logloss', 'auc'},\n",
    "    'metric_freq': 1,\n",
    "    'is_training_metric': True,\n",
    "    'max_bin': 255,\n",
    "    'learning_rate': 0.1,\n",
    "    'num_leaves': 63,\n",
    "    'tree_learner': 'serial',\n",
    "    'feature_fraction': 0.8,\n",
    "    'bagging_fraction': 0.8,\n",
    "    'bagging_freq': 5,\n",
    "    'min_data_in_leaf': 50,\n",
    "    'min_sum_hessian_in_leaf': 5,\n",
    "    'is_enable_sparse': True,\n",
    "    'use_two_round_loading': False,\n",
    "    'is_save_binary_file': False,\n",
    "    'output_model': 'LightGBM_model.txt',\n",
    "    'num_machines': 1,\n",
    "    'local_listen_port': 12400,\n",
    "    'machine_list_file': 'mlist.txt',\n",
    "    'verbose': 0,\n",
    "    'subsample_for_bin': 200000,\n",
    "    'min_child_samples': 20,\n",
    "    'min_child_weight': 0.001,\n",
    "    'min_split_gain': 0.0,\n",
    "    'colsample_bytree': 1.0,\n",
    "    'reg_alpha': 0.0,\n",
    "    'reg_lambda': 0.0\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading data...\n",
      "(7000, 29)\n",
      "(500, 29)\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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>19</th>\n",
       "      <th>20</th>\n",
       "      <th>21</th>\n",
       "      <th>22</th>\n",
       "      <th>23</th>\n",
       "      <th>24</th>\n",
       "      <th>25</th>\n",
       "      <th>26</th>\n",
       "      <th>27</th>\n",
       "      <th>28</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0.869</td>\n",
       "      <td>-0.635</td>\n",
       "      <td>0.226</td>\n",
       "      <td>0.327</td>\n",
       "      <td>-0.690</td>\n",
       "      <td>0.754</td>\n",
       "      <td>-0.249</td>\n",
       "      <td>-1.092</td>\n",
       "      <td>0.000</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.010</td>\n",
       "      <td>-0.046</td>\n",
       "      <td>3.102</td>\n",
       "      <td>1.354</td>\n",
       "      <td>0.980</td>\n",
       "      <td>0.978</td>\n",
       "      <td>0.920</td>\n",
       "      <td>0.722</td>\n",
       "      <td>0.989</td>\n",
       "      <td>0.877</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0.908</td>\n",
       "      <td>0.329</td>\n",
       "      <td>0.359</td>\n",
       "      <td>1.498</td>\n",
       "      <td>-0.313</td>\n",
       "      <td>1.096</td>\n",
       "      <td>-0.558</td>\n",
       "      <td>-1.588</td>\n",
       "      <td>2.173</td>\n",
       "      <td>...</td>\n",
       "      <td>-1.139</td>\n",
       "      <td>-0.001</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.302</td>\n",
       "      <td>0.833</td>\n",
       "      <td>0.986</td>\n",
       "      <td>0.978</td>\n",
       "      <td>0.780</td>\n",
       "      <td>0.992</td>\n",
       "      <td>0.798</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>0.799</td>\n",
       "      <td>1.471</td>\n",
       "      <td>-1.636</td>\n",
       "      <td>0.454</td>\n",
       "      <td>0.426</td>\n",
       "      <td>1.105</td>\n",
       "      <td>1.282</td>\n",
       "      <td>1.382</td>\n",
       "      <td>0.000</td>\n",
       "      <td>...</td>\n",
       "      <td>1.129</td>\n",
       "      <td>0.900</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.910</td>\n",
       "      <td>1.108</td>\n",
       "      <td>0.986</td>\n",
       "      <td>0.951</td>\n",
       "      <td>0.803</td>\n",
       "      <td>0.866</td>\n",
       "      <td>0.780</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>1.344</td>\n",
       "      <td>-0.877</td>\n",
       "      <td>0.936</td>\n",
       "      <td>1.992</td>\n",
       "      <td>0.882</td>\n",
       "      <td>1.786</td>\n",
       "      <td>-1.647</td>\n",
       "      <td>-0.942</td>\n",
       "      <td>0.000</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.678</td>\n",
       "      <td>-1.360</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.947</td>\n",
       "      <td>1.029</td>\n",
       "      <td>0.999</td>\n",
       "      <td>0.728</td>\n",
       "      <td>0.869</td>\n",
       "      <td>1.027</td>\n",
       "      <td>0.958</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>1.105</td>\n",
       "      <td>0.321</td>\n",
       "      <td>1.522</td>\n",
       "      <td>0.883</td>\n",
       "      <td>-1.205</td>\n",
       "      <td>0.681</td>\n",
       "      <td>-1.070</td>\n",
       "      <td>-0.922</td>\n",
       "      <td>0.000</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.374</td>\n",
       "      <td>0.113</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.756</td>\n",
       "      <td>1.361</td>\n",
       "      <td>0.987</td>\n",
       "      <td>0.838</td>\n",
       "      <td>1.133</td>\n",
       "      <td>0.872</td>\n",
       "      <td>0.808</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 29 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   0      1      2      3      4      5      6      7      8      9   ...    \\\n",
       "0   1  0.869 -0.635  0.226  0.327 -0.690  0.754 -0.249 -1.092  0.000  ...     \n",
       "1   1  0.908  0.329  0.359  1.498 -0.313  1.096 -0.558 -1.588  2.173  ...     \n",
       "2   1  0.799  1.471 -1.636  0.454  0.426  1.105  1.282  1.382  0.000  ...     \n",
       "3   0  1.344 -0.877  0.936  1.992  0.882  1.786 -1.647 -0.942  0.000  ...     \n",
       "4   1  1.105  0.321  1.522  0.883 -1.205  0.681 -1.070 -0.922  0.000  ...     \n",
       "\n",
       "      19     20     21     22     23     24     25     26     27     28  \n",
       "0 -0.010 -0.046  3.102  1.354  0.980  0.978  0.920  0.722  0.989  0.877  \n",
       "1 -1.139 -0.001  0.000  0.302  0.833  0.986  0.978  0.780  0.992  0.798  \n",
       "2  1.129  0.900  0.000  0.910  1.108  0.986  0.951  0.803  0.866  0.780  \n",
       "3 -0.678 -1.360  0.000  0.947  1.029  0.999  0.728  0.869  1.027  0.958  \n",
       "4 -0.374  0.113  0.000  0.756  1.361  0.987  0.838  1.133  0.872  0.808  \n",
       "\n",
       "[5 rows x 29 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(\"Loading data...\")\n",
    "df_train=pd.read_csv('files/data/python83/regression.train.txt',header=None,sep='\\t')\n",
    "df_test=pd.read_csv('files/data/python83/regression.test.txt',header=None,sep='\\t')\n",
    "print(df_train.shape)\n",
    "print(df_test.shape)\n",
    "df_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train=df_train[0].values\n",
    "y_test=df_test[0].values\n",
    "X_train=df_train.drop(0,axis=1).values\n",
    "X_test=df_test.drop(0,axis=1).values\n",
    "lgb_train = lgb.Dataset(X_train, y_train)\n",
    "lgb_eval = lgb.Dataset(X_test, y_test, reference=lgb_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1]\tvalid_0's auc: 0.718121\tvalid_0's binary_logloss: 0.671716\n",
      "[2]\tvalid_0's auc: 0.734415\tvalid_0's binary_logloss: 0.658775\n",
      "[3]\tvalid_0's auc: 0.767592\tvalid_0's binary_logloss: 0.644696\n",
      "[4]\tvalid_0's auc: 0.774332\tvalid_0's binary_logloss: 0.63497\n",
      "[5]\tvalid_0's auc: 0.780597\tvalid_0's binary_logloss: 0.622594\n",
      "[6]\tvalid_0's auc: 0.782298\tvalid_0's binary_logloss: 0.613867\n",
      "[7]\tvalid_0's auc: 0.785942\tvalid_0's binary_logloss: 0.606095\n",
      "[8]\tvalid_0's auc: 0.787587\tvalid_0's binary_logloss: 0.598635\n",
      "[9]\tvalid_0's auc: 0.790361\tvalid_0's binary_logloss: 0.59144\n",
      "[10]\tvalid_0's auc: 0.794521\tvalid_0's binary_logloss: 0.58343\n",
      "[11]\tvalid_0's auc: 0.798842\tvalid_0's binary_logloss: 0.576479\n",
      "[12]\tvalid_0's auc: 0.798681\tvalid_0's binary_logloss: 0.572351\n",
      "[13]\tvalid_0's auc: 0.798826\tvalid_0's binary_logloss: 0.567695\n",
      "[14]\tvalid_0's auc: 0.80089\tvalid_0's binary_logloss: 0.564266\n",
      "[15]\tvalid_0's auc: 0.803196\tvalid_0's binary_logloss: 0.559937\n",
      "[16]\tvalid_0's auc: 0.804373\tvalid_0's binary_logloss: 0.557184\n",
      "[17]\tvalid_0's auc: 0.808759\tvalid_0's binary_logloss: 0.55253\n",
      "[18]\tvalid_0's auc: 0.811452\tvalid_0's binary_logloss: 0.549222\n",
      "[19]\tvalid_0's auc: 0.811113\tvalid_0's binary_logloss: 0.548561\n",
      "[20]\tvalid_0's auc: 0.813725\tvalid_0's binary_logloss: 0.544424\n",
      "[21]\tvalid_0's auc: 0.814822\tvalid_0's binary_logloss: 0.541876\n",
      "[22]\tvalid_0's auc: 0.818257\tvalid_0's binary_logloss: 0.538312\n",
      "[23]\tvalid_0's auc: 0.818386\tvalid_0's binary_logloss: 0.536112\n",
      "[24]\tvalid_0's auc: 0.817499\tvalid_0's binary_logloss: 0.535519\n",
      "[25]\tvalid_0's auc: 0.819063\tvalid_0's binary_logloss: 0.533311\n",
      "[26]\tvalid_0's auc: 0.821723\tvalid_0's binary_logloss: 0.530656\n",
      "[27]\tvalid_0's auc: 0.820224\tvalid_0's binary_logloss: 0.530679\n",
      "[28]\tvalid_0's auc: 0.820675\tvalid_0's binary_logloss: 0.52913\n",
      "[29]\tvalid_0's auc: 0.818773\tvalid_0's binary_logloss: 0.52929\n",
      "[30]\tvalid_0's auc: 0.82103\tvalid_0's binary_logloss: 0.527165\n",
      "[31]\tvalid_0's auc: 0.821514\tvalid_0's binary_logloss: 0.52574\n",
      "[32]\tvalid_0's auc: 0.822884\tvalid_0's binary_logloss: 0.524064\n",
      "[33]\tvalid_0's auc: 0.825819\tvalid_0's binary_logloss: 0.52034\n",
      "[34]\tvalid_0's auc: 0.825561\tvalid_0's binary_logloss: 0.519583\n",
      "[35]\tvalid_0's auc: 0.823755\tvalid_0's binary_logloss: 0.519949\n",
      "[36]\tvalid_0's auc: 0.825464\tvalid_0's binary_logloss: 0.518062\n",
      "[37]\tvalid_0's auc: 0.82698\tvalid_0's binary_logloss: 0.516739\n",
      "[38]\tvalid_0's auc: 0.827174\tvalid_0's binary_logloss: 0.516238\n",
      "[39]\tvalid_0's auc: 0.827174\tvalid_0's binary_logloss: 0.515569\n",
      "[40]\tvalid_0's auc: 0.828302\tvalid_0's binary_logloss: 0.514191\n",
      "[41]\tvalid_0's auc: 0.829125\tvalid_0's binary_logloss: 0.513131\n",
      "[42]\tvalid_0's auc: 0.829496\tvalid_0's binary_logloss: 0.512421\n",
      "[43]\tvalid_0's auc: 0.830463\tvalid_0's binary_logloss: 0.510758\n",
      "[44]\tvalid_0's auc: 0.831431\tvalid_0's binary_logloss: 0.509956\n",
      "[45]\tvalid_0's auc: 0.832059\tvalid_0's binary_logloss: 0.509556\n",
      "[46]\tvalid_0's auc: 0.831543\tvalid_0's binary_logloss: 0.509047\n",
      "[47]\tvalid_0's auc: 0.831189\tvalid_0's binary_logloss: 0.509304\n",
      "[48]\tvalid_0's auc: 0.830657\tvalid_0's binary_logloss: 0.509256\n",
      "[49]\tvalid_0's auc: 0.831011\tvalid_0's binary_logloss: 0.508343\n",
      "[50]\tvalid_0's auc: 0.831334\tvalid_0's binary_logloss: 0.507408\n",
      "[51]\tvalid_0's auc: 0.833124\tvalid_0's binary_logloss: 0.504707\n",
      "[52]\tvalid_0's auc: 0.834527\tvalid_0's binary_logloss: 0.502826\n",
      "[53]\tvalid_0's auc: 0.834817\tvalid_0's binary_logloss: 0.502681\n",
      "[54]\tvalid_0's auc: 0.835059\tvalid_0's binary_logloss: 0.502092\n",
      "[55]\tvalid_0's auc: 0.834043\tvalid_0's binary_logloss: 0.503374\n",
      "[56]\tvalid_0's auc: 0.833511\tvalid_0's binary_logloss: 0.503293\n",
      "[57]\tvalid_0's auc: 0.835994\tvalid_0's binary_logloss: 0.501813\n",
      "[58]\tvalid_0's auc: 0.834333\tvalid_0's binary_logloss: 0.50282\n",
      "[59]\tvalid_0's auc: 0.832446\tvalid_0's binary_logloss: 0.504135\n",
      "[60]\tvalid_0's auc: 0.832575\tvalid_0's binary_logloss: 0.504659\n",
      "[61]\tvalid_0's auc: 0.831785\tvalid_0's binary_logloss: 0.504071\n",
      "[62]\tvalid_0's auc: 0.831914\tvalid_0's binary_logloss: 0.504557\n",
      "[63]\tvalid_0's auc: 0.832301\tvalid_0's binary_logloss: 0.50366\n",
      "[64]\tvalid_0's auc: 0.832833\tvalid_0's binary_logloss: 0.50341\n",
      "[65]\tvalid_0's auc: 0.83214\tvalid_0's binary_logloss: 0.503906\n",
      "[66]\tvalid_0's auc: 0.832188\tvalid_0's binary_logloss: 0.503832\n",
      "[67]\tvalid_0's auc: 0.833059\tvalid_0's binary_logloss: 0.503122\n",
      "[68]\tvalid_0's auc: 0.83243\tvalid_0's binary_logloss: 0.503605\n",
      "[69]\tvalid_0's auc: 0.832704\tvalid_0's binary_logloss: 0.502911\n",
      "[70]\tvalid_0's auc: 0.833172\tvalid_0's binary_logloss: 0.502656\n",
      "[71]\tvalid_0's auc: 0.834252\tvalid_0's binary_logloss: 0.501422\n",
      "[72]\tvalid_0's auc: 0.833333\tvalid_0's binary_logloss: 0.502256\n",
      "[73]\tvalid_0's auc: 0.832414\tvalid_0's binary_logloss: 0.503688\n",
      "[74]\tvalid_0's auc: 0.83156\tvalid_0's binary_logloss: 0.504318\n",
      "[75]\tvalid_0's auc: 0.831237\tvalid_0's binary_logloss: 0.504458\n",
      "[76]\tvalid_0's auc: 0.832753\tvalid_0's binary_logloss: 0.502694\n",
      "[77]\tvalid_0's auc: 0.830318\tvalid_0's binary_logloss: 0.504696\n",
      "[78]\tvalid_0's auc: 0.830931\tvalid_0's binary_logloss: 0.504329\n",
      "[79]\tvalid_0's auc: 0.829641\tvalid_0's binary_logloss: 0.505392\n",
      "[80]\tvalid_0's auc: 0.82898\tvalid_0's binary_logloss: 0.506031\n",
      "[81]\tvalid_0's auc: 0.828528\tvalid_0's binary_logloss: 0.506144\n",
      "[82]\tvalid_0's auc: 0.828109\tvalid_0's binary_logloss: 0.506959\n",
      "[83]\tvalid_0's auc: 0.830624\tvalid_0's binary_logloss: 0.504479\n",
      "[84]\tvalid_0's auc: 0.832446\tvalid_0's binary_logloss: 0.502562\n",
      "[85]\tvalid_0's auc: 0.830882\tvalid_0's binary_logloss: 0.50385\n",
      "[86]\tvalid_0's auc: 0.830882\tvalid_0's binary_logloss: 0.504381\n",
      "[87]\tvalid_0's auc: 0.830882\tvalid_0's binary_logloss: 0.50378\n",
      "[88]\tvalid_0's auc: 0.829673\tvalid_0's binary_logloss: 0.504868\n",
      "[89]\tvalid_0's auc: 0.830592\tvalid_0's binary_logloss: 0.503496\n",
      "[90]\tvalid_0's auc: 0.828963\tvalid_0's binary_logloss: 0.50518\n",
      "[91]\tvalid_0's auc: 0.828609\tvalid_0's binary_logloss: 0.505088\n",
      "[92]\tvalid_0's auc: 0.827932\tvalid_0's binary_logloss: 0.505781\n",
      "[93]\tvalid_0's auc: 0.827319\tvalid_0's binary_logloss: 0.505743\n",
      "[94]\tvalid_0's auc: 0.825303\tvalid_0's binary_logloss: 0.507553\n",
      "[95]\tvalid_0's auc: 0.824739\tvalid_0's binary_logloss: 0.50872\n",
      "[96]\tvalid_0's auc: 0.824416\tvalid_0's binary_logloss: 0.509107\n",
      "[97]\tvalid_0's auc: 0.824561\tvalid_0's binary_logloss: 0.509283\n",
      "[98]\tvalid_0's auc: 0.824997\tvalid_0's binary_logloss: 0.508737\n",
      "[99]\tvalid_0's auc: 0.823933\tvalid_0's binary_logloss: 0.510582\n",
      "[100]\tvalid_0's auc: 0.823642\tvalid_0's binary_logloss: 0.51086\n"
     ]
    }
   ],
   "source": [
    "gbm = lgb.train(params,\n",
    "                lgb_train,\n",
    "                valid_sets=lgb_eval)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "\n",
    "gridParams = {\n",
    "    'learning_rate': [ 0.1],\n",
    "    'num_leaves': [63],\n",
    "    'boosting_type' : ['gbdt'],\n",
    "    'objective' : ['binary']\n",
    "}\n",
    "\n",
    "mdl = lgb.LGBMClassifier(\n",
    "    task = params['task'],\n",
    "    metric = params['metric'],\n",
    "    metric_freq = params['metric_freq'],\n",
    "    is_training_metric = params['is_training_metric'],\n",
    "    max_bin = params['max_bin'],\n",
    "    tree_learner = params['tree_learner'],\n",
    "    feature_fraction = params['feature_fraction'],\n",
    "    bagging_fraction = params['bagging_fraction'],\n",
    "    bagging_freq = params['bagging_freq'],\n",
    "    min_data_in_leaf = params['min_data_in_leaf'],\n",
    "    min_sum_hessian_in_leaf = params['min_sum_hessian_in_leaf'],\n",
    "    is_enable_sparse = params['is_enable_sparse'],\n",
    "    use_two_round_loading = params['use_two_round_loading'],\n",
    "    is_save_binary_file = params['is_save_binary_file'],\n",
    "    n_jobs = -1\n",
    ")\n",
    "\n",
    "scoring = {'AUC': 'roc_auc'}\n",
    "\n",
    "# Create the grid\n",
    "grid = GridSearchCV(mdl, gridParams, verbose=2, cv=5, scoring=scoring, n_jobs=-1, refit='AUC')\n",
    "# Run the grid\n",
    "grid.fit(X_train, y_train)\n",
    "\n",
    "print('Best parameters found by grid search are:', grid.best_params_)\n",
    "print('Best score found by grid search is:', grid.best_score_)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 实例2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>21</th>\n",
       "      <th>22</th>\n",
       "      <th>23</th>\n",
       "      <th>24</th>\n",
       "      <th>25</th>\n",
       "      <th>26</th>\n",
       "      <th>27</th>\n",
       "      <th>28</th>\n",
       "      <th>29</th>\n",
       "      <th>30</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>...</td>\n",
       "      <td>0.199728</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>-0.333333</td>\n",
       "      <td>-0.65</td>\n",
       "      <td>-0.621622</td>\n",
       "      <td>-0.444444</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>-0.176471</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>4</td>\n",
       "      <td>3</td>\n",
       "      <td>...</td>\n",
       "      <td>0.060483</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>-0.333333</td>\n",
       "      <td>0.65</td>\n",
       "      <td>-0.729730</td>\n",
       "      <td>-0.555556</td>\n",
       "      <td>-0.866667</td>\n",
       "      <td>-0.529412</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "      <td>...</td>\n",
       "      <td>0.900553</td>\n",
       "      <td>1.0</td>\n",
       "      <td>-1.0</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>-0.55</td>\n",
       "      <td>-0.513514</td>\n",
       "      <td>-0.222222</td>\n",
       "      <td>-1.000000</td>\n",
       "      <td>-0.294118</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3 rows × 31 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   0  1  2  3  4  5  6  7  8  9 ...        21   22   23        24    25  \\\n",
       "0  2  1  4  1  1  1  2  2  4  3 ...  0.199728 -1.0 -1.0 -0.333333 -0.65   \n",
       "1  1  1  4  1  4  0  3  3  4  3 ...  0.060483 -1.0 -1.0 -0.333333  0.65   \n",
       "2  1  1  3  1  1  1  1  2  2  3 ...  0.900553  1.0 -1.0 -1.000000 -0.55   \n",
       "\n",
       "         26        27        28        29  30  \n",
       "0 -0.621622 -0.444444 -1.000000 -0.176471   0  \n",
       "1 -0.729730 -0.555556 -0.866667 -0.529412   0  \n",
       "2 -0.513514 -0.222222 -1.000000 -0.294118   1  \n",
       "\n",
       "[3 rows x 31 columns]"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import lightgbm as lgb\n",
    "from sklearn.grid_search import GridSearchCV\n",
    "from sklearn.model_selection import train_test_split\n",
    "train_data=pd.read_csv('files/data/python87/train.csv')\n",
    "train_data.head(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "y=train_data.pop('30').values # 用pop方式将训练数据中的标签值y取出来，作为训练目标，这里的‘30’是标签的列名"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "col = train_data.columns   \n",
    "x = train_data[col].values  # 剩下的列作为训练数据\n",
    "train_x, valid_x, train_y, valid_y = train_test_split(x, y, test_size=0.333, random_state=0)   # 分训练集和验证集\n",
    "train = lgb.Dataset(train_x, train_y)\n",
    "valid = lgb.Dataset(valid_x, valid_y, reference=train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 需要优化的参数\n",
    "parameters = {\n",
    "              'max_depth': [15, 20, 25, 30, 35],\n",
    "              'learning_rate': [0.01, 0.02, 0.05, 0.1, 0.15],\n",
    "              'feature_fraction': [0.6, 0.7, 0.8, 0.9, 0.95],\n",
    "#               'bagging_fraction': [0.6, 0.7, 0.8, 0.9, 0.95],\n",
    "#               'bagging_freq': [2, 4, 5, 6, 8],\n",
    "#               'lambda_l1': [0, 0.1, 0.4, 0.5, 0.6],\n",
    "#               'lambda_l2': [0, 10, 15, 35, 40],\n",
    "#               'cat_smooth': [1, 10, 15, 20, 35]\n",
    "}\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "gbm = lgb.LGBMClassifier(boosting_type='gbdt',\n",
    "                         objective = 'binary',\n",
    "                         metric = 'auc',\n",
    "                         verbose = 0,\n",
    "                         learning_rate = 0.01,\n",
    "                         num_leaves = 35,\n",
    "                         feature_fraction=0.8,\n",
    "                         bagging_fraction= 0.9,\n",
    "                         bagging_freq= 8,\n",
    "                         lambda_l1= 0.6,\n",
    "                         lambda_l2= 0)\n",
    "# 有了gridsearch我们便不需要fit函数\n",
    "gsearch = GridSearchCV(gbm, param_grid=parameters, scoring='accuracy', cv=3)\n",
    "gsearch.fit(train_x, train_y)\n",
    "\n",
    "print(\"Best score: %0.3f\" % gsearch.best_score_)\n",
    "print(\"Best parameters set:\")\n",
    "best_parameters = gsearch.best_estimator_.get_params()\n",
    "for param_name in sorted(parameters.keys()):\n",
    "    print(\"\\t%s: %r\" % (param_name, best_parameters[param_name]))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "https://juejin.im/post/5b76437ae51d45666b5d9b05"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}