Lightgbm支持两种形式的调用接口：原生形式和sklearn接口的形式。

原生形式

sklearn接口的形式

1. 导入包

import lightgbm as lgb
from sklearn.metrics import mean_squared_error
from sklearn.datasets import load_boston
from sklearn.model_selection import train_test_split

2.加载boston房价数据

boston = load_boston()
data = boston.data
target = boston.target

3.切分数据集

X_train, X_test, y_train, y_test = train_test_split(data, target, test_size=0.2)

4.创建成lgb特征的数据集格式

lgb_train = lgb.Dataset(X_train, y_train)
#If this is Dataset for validation, training data should be used as reference.
lgb_eval = lgb.Dataset(X_test, y_test, reference=lgb_train) 

5.将参数写成字典下形式(Boston房价问题为回归问题，objective设置为regression)

params = {
    'task': 'train',
    'boosting_type': 'gbdt',  # 设置提升类型
    'objective': 'regression',  # 目标函数
    'metric': {'l2', 'auc'},  # 评估函数
    'num_leaves': 31,  # 叶子节点数
    'learning_rate': 0.05,  # 学习速率
    'feature_fraction': 0.9,  # 建树的特征选择比例
    'bagging_fraction': 0.8,  # 建树的样本采样比例
    'bagging_freq': 5,  # k 意味着每 k 次迭代执行bagging
    'verbose': 1  # <0 显示致命的, =0 显示错误 (警告), >0 显示信息
}

6.交叉验证（CrossValidation）及训练

• num_boost_round : int, optional (default=100)，Number of boosting iterations.

gbm = lgb.train(params, lgb_train, num_boost_round=20, valid_sets=lgb_eval, early_stopping_rounds=5)

7.保存模型到文件

import joblib
#gbm.save_model('model.txt')
joblib.dump(lgb, '.lgb.pkl')

8.预测数据集

y_pred = gbm.predict(X_test, num_iteration=gbm.best_iteration)

9.评估模型

print('The MSE of prediction is:', mean_squared_error(y_test, y_pred) ** 0.5)

sklearn接口的形式

1.加载数据

boston = load_boston()
data = boston.data
target = boston.target
X_train, X_test, y_train, y_test = train_test_split(data, target, test_size=0.2)

2.创建模型，训练模型

gbm = lgb.LGBMRegressor(objective='regression', num_leaves=31, learning_rate=0.05, n_estimators=20)
gbm.fit(X_train, y_train, eval_set=[(X_test, y_test)], eval_metric='l1', early_stopping_rounds=5)

3.模型预测

y_pred = gbm.predict(X_test, num_iteration=gbm.best_iteration_)

4.模型评估

print('The MSE of prediction is:', mean_squared_error(y_test, y_pred) ** 0.5)

5.特征重要性

print('Feature importances:', list(gbm.feature_importances_))

6.网格搜索，参数优化

from sklearn.model_selection import GridSearchCV
estimator = lgb.LGBMRegressor(num_leaves=31)
param_grid = {
    'learning_rate': [0.01, 0.1, 1],
    'n_estimators': [20, 40]
}
gbm = GridSearchCV(estimator, param_grid)
gbm.fit(X_train, y_train)
print('Best parameters found by grid search are:', gbm.best_params_)