from sklearn.datasets import load_irisfrom sklearn.ensemble import RandomForestClassifier, StackingClassifierfrom sklearn.linear_model import LogisticRegressionfrom sklearn.model_selection import train_test_splitfrom sklearn.neighbors import KNeighborsClassifierX, y = load_iris(return_X_y=True)# 创建基础学习器base_learners = [ ('rf_1', RandomForestClassifier(n_estimators=10, random_state=42)), ('rf_2', KNeighborsClassifier(n_neighbors=5)) ]# 初始化带有元学习器的堆叠分类器clf = StackingClassifier(estimators=base_learners, final_estimator=LogisticRegression())如何查看clf中的堆叠系数（即逻辑回归系数）？根据这个帖子的建议，我尝试了以下方法，但没有成功：
pipeline = joblib.load('clf')#第一个模型的系数，遍历estimators_获取其余的pipeline['stackingclassifier'].estimators_[0].coef_回答：
总结
你忘了拟合分类器以获取它们的属性
工作示例：
from sklearn.datasets import load_irisfrom sklearn.ensemble import RandomForestClassifier, StackingClassifierfrom sklearn.linear_model import LogisticRegressionfrom sklearn.neighbors import KNeighborsClassifierX, y = load_iris(return_X_y=True)# 创建基础学习器base_learners = [ ('rf_1', RandomForestClassifier(n_estimators=10, random_state=42)), ('rf_2', KNeighborsClassifier(n_neighbors=5))]# 初始化带有元学习器的堆叠分类器clf = StackingClassifier(estimators=base_learners, stack_method='predict', final_estimator=LogisticRegression())# 你忘了拟合分类器以获取属性clf.fit(X, y)print(clf.final_estimator_.coef_) # 逻辑回归系数# RandomForestClassifier .. 你可以通过.name_of_attribute访问它的属性print(clf.named_estimators_['rf_1'])# KNeighborsClassifier .. 你可以通过.name_of_attribute访问它的属性print(clf.named_estimators_['rf_2'])print(clf.estimators_) # 列表中的所有估计器输出：
[[-2.68962706 -2.69561738] [ 0.22543939 -0.01107915] [ 2.46418768 2.70669652]]RandomForestClassifier(n_estimators=10, random_state=42)KNeighborsClassifier()[RandomForestClassifier(n_estimators=10, random_state=42), KNeighborsClassifier()]还要注意根据你的需求更改stack_method参数 1。
