light gbm

lgbm.py

import lightgbm as lgb
from sklearn.model_selection import GridSearchCV

param_grid = {
    'objective': ["binary"],
    'metric': ['binary_logloss'],
    'boosting': ['gbdt','dart'],
    'lambda_l1': [0, 0.1, 0.01],
    'bagging_fraction': [0.7, 0.9],
    'bagging_freq': [0,1],
    'num_leaves': [32, 64, 128],
    'max_depth': [-1, 10],
    'max_bin': [100, 255],
    'num_iterations': [100, 200],
    'learning_rate':[0.01,0.1],
    'min_data_in_leaf': [3,5]
}

estimator = lgb.LGBMClassifier()

searcher = GridSearchCV(
        estimator=estimator,
        param_grid=param_grid, 
        cv=5, 
        n_jobs=-1, 
        scoring='accuracy',
        verbose=2)

grid_model = searcher.fit(X_train, y_train)

y_pred =grid_model.predict(X_test)