python决策树可视化_决策树可视化-python

决策树相较于其他机器学习模型具有较好的解释性，也容易将其结果进行可视化展示，python中sklearn.tree的export_graphviz函数可以将决策树结果以doc或dot文件的形式输出，然后将决策树结果可视化，下面举个简单例子。

1、创建数据集from sklearn.datasets import make_classification

X, y = make_classification(n_samples=10000,

n_features=10,

n_informative=5,

n_redundant=0,

n_repeated=0,

n_classes=2,

weights=[0.8, 0.2],

flip_y=0.01,

random_state=1234)

2、训练决策树from sklearn.tree import DecisionTreeClassifier

Dtree = DecisionTreeClassifier(criterion='gini',

max_depth=3,

min_samples_split=int(0.03 * X.shape[0]),

min_samples_leaf=int(0.01 * X.shape[0]),

random_state=0)

Dtree.fit(X, y)

3、输出决策树结果文件from sklearn.tree import export_graphviz

export_graphviz(Dtree,

out_file='D:\\study\\Dtree.dot')

在'D:\study'中生成了'Dt