使用 lightgbm 的特征重要性

Question

我正在尝试运行我的 lightgbm 进行功能选择，如下所示；初始化# Initialize an empty array to hold feature importancesfeature_importances = np.zeros(features_sample.shape[1])# Create the model with several hyperparametersmodel = lgb.LGBMClassifier(objective='binary',          boosting_type = 'goss',          n_estimators = 10000, class_weight ='balanced')然后我适合模型如下# Fit the model twice to avoid overfittingfor i in range(2):   # Split into training and validation set   train_features, valid_features, train_y, valid_y = train_test_split(train_X, train_Y, test_size = 0.25, random_state = i)   # Train using early stopping   model.fit(train_features, train_y, early_stopping_rounds=100, eval_set = [(valid_features, valid_y)],              eval_metric = 'auc', verbose = 200)   # Record the feature importances   feature_importances += model.feature_importances_但我收到以下错误Training until validation scores don't improve for 100 rounds. Early stopping, best iteration is: [6]  valid_0's auc: 0.88648ValueError: operands could not be broadcast together with shapes (87,) (83,) (87,) 

Answer 1

根据我们是训练modelusingscikit-learn还是lightgbm方法，为了获得重要性，我们应该分别选择feature_importances_属性或feature_importance()函数，就像在这个例子中一样（其中model是lgbm.fit() / lgbm.train(), 和的结果train_columns = x_train_df.columns）：

import pandas as pd

def get_lgbm_varimp(model, train_columns, max_vars=50):

    

    if "basic.Booster" in str(model.__class__):

        # lightgbm.basic.Booster was trained directly, so using feature_importance() function 

        cv_varimp_df = pd.DataFrame([train_columns, model.feature_importance()]).T

    else:

        # Scikit-learn API LGBMClassifier or LGBMRegressor was fitted, 

        # so using feature_importances_ property

        cv_varimp_df = pd.DataFrame([train_columns, model.feature_importances_]).T

    cv_varimp_df.columns = ['feature_name', 'varimp']

    cv_varimp_df.sort_values(by='varimp', ascending=False, inplace=True)

    cv_varimp_df = cv_varimp_df.iloc[0:max_vars]   

    return cv_varimp_df

    

请注意，我们依赖于这样一个假设，即特征重要性值的排序就像训练期间模型矩阵列的排序（包括 one-hot dummy cols），请参阅LightGBM #209。