在使用scikit learn中的DecisionTreeClassifier时，获取决策树和重要特征是非常容易的。然而，如果我使用了bagging函数，例如BaggingClassifier，我就无法获取这些信息了。

由于我们需要使用BaggingClassifier来拟合模型，我无法返回与DecisionTreeClassifier相关的任何结果（打印树（图形）、feature_importances_等）。

这是我的脚本：

seed = 7n_iterations = 199DTC = DecisionTreeClassifier(random_state=seed,                                                 max_depth=None,                                                 min_impurity_split= 0.2,                                                 min_samples_leaf=6,                                                 max_features=None, #如果为None，则max_features=n_features.                                                 max_leaf_nodes=20,                                                 criterion='gini',                                                 splitter='best',                                                 )#parametersDTC = {'max_depth':range(3,10), 'max_leaf_nodes':range(10, 30)}parameters = {'max_features':range(1,200)}dt = RandomizedSearchCV(BaggingClassifier(base_estimator=DTC,                              #max_samples=1,                              n_estimators=100,                              #max_features=1,                              bootstrap = False,                              bootstrap_features = True, random_state=seed),                        parameters, n_iter=n_iterations, n_jobs=14, cv=kfold,                        error_score='raise', random_state=seed, refit=True) #min_samples_leaf=10# 拟合模型fit_dt= dt.fit(X_train, Y_train)print(dir(fit_dt))tree_model = dt.best_estimator_# 打印重要特征（无法工作）features = tree_model.feature_importances_print(features)rank = np.argsort(features)[::-1]print(rank[:12])print(sorted(list(zip(features))))# 导入图像（无法工作）from sklearn.externals.six import StringIOtree.export_graphviz(dt.best_estimator_, out_file='tree.dot') # 必要的步骤以绘制图形dot_data = StringIO() # 需要理解，但可能与读取字符串有关tree.export_graphviz(dt.best_estimator_, out_file=dot_data, filled=True, class_names= target_names, rounded=True, special_characters=True)graph = pydotplus.graph_from_dot_data(dot_data.getvalue())img = Image(graph.create_png())print(dir(img)) # 使用dir我们可以检查graph.create_png的可能性with open("my_tree.png", "wb") as png:    png.write(img.data)

我得到了类似于’BaggingClassifier’ object has no attribute ‘tree_’ 和 ‘BaggingClassifier’ object has no attribute ‘feature_importances’ 的错误。有人知道我如何获取这些信息吗？谢谢。

回答：

根据文档，BaggingClassifier对象确实没有’feature_importances’属性。你仍然可以按照这个问题的答案中描述的那样自己计算它：Feature importances – Bagging, scikit-learn

你可以使用属性estimators_访问在BaggingClassifier拟合过程中生成的树，如下例所示：

from sklearn import svm, datasetsfrom sklearn.model_selection import GridSearchCVfrom sklearn.ensemble import BaggingClassifieriris = datasets.load_iris()clf = BaggingClassifier(n_estimators=3)clf.fit(iris.data, iris.target)clf.estimators_

clf.estimators_是一个包含3个拟合决策树的列表：

[DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=None,             max_features=None, max_leaf_nodes=None,             min_impurity_split=1e-07, min_samples_leaf=1,             min_samples_split=2, min_weight_fraction_leaf=0.0,             presort=False, random_state=1422640898, splitter='best'), DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=None,             max_features=None, max_leaf_nodes=None,             min_impurity_split=1e-07, min_samples_leaf=1,             min_samples_split=2, min_weight_fraction_leaf=0.0,             presort=False, random_state=1968165419, splitter='best'), DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=None,             max_features=None, max_leaf_nodes=None,             min_impurity_split=1e-07, min_samples_leaf=1,             min_samples_split=2, min_weight_fraction_leaf=0.0,             presort=False, random_state=2103976874, splitter='best')]

所以你可以遍历这个列表并访问每棵树。