基于sklearn的 BaseEstimator开发接口：模型融合Stacking

转载：https://github.com/LearningFromBest/CMB-credit-card-department-prediction-of-purchasing-behavior-in-consumer-finance-scenario/blob/master/stacking.py

from sklearn.base import BaseEstimator, TransformerMixin, RegressorMixin, clone
from sklearn.model_selection import KFold, cross_val_score, train_test_split
import numpy as np

class Stacking(BaseEstimator):
 
    def __init__(self,base_models,meta_model,folds=5):
        self.base_models = base_models
        self.meta_models = meta_model
        self.folds = folds

    def fit(self,X,y):
         self.base_models_ = [[] for x in self.base_models]
         self.meta_models_ = clone(self.meta_models)
         kfold = KFold(n_splits=self.folds,shuffle=True,random_state=156)

         #用于存储每一折每个估计器的输出，拼接成行的特征
         matrix_out_fit = np.zeros(X.shape[0],len(self.base_models))

         for i , model in enumerate(self.base_models):
             for train_index , test_index in kfold.split(X,y):
                 temp_model = clone(model)
                 self.base_models_[i].append(temp_model)
                 temp_model.fit(X[train_index],y[train_index])
                 matrix_out_fit[test_index,i] = temp_model.predict(X[test_index])

         self.meta_models_.fit(matrix_out_fit,y)

         return self

    def predict(self,X):
        meta_features = np.column_stack([np.column_stack([model.predict(X) for model in base_models]).mean(axis=1) for base_models in self.base_models_])
        return self.meta_model_.predict(meta_features)