scikit-learn实现决策树分类与回归以及调参说明

落日映苍穹つ 2024-04-19 16:11 59阅读 0赞

# 分类： #

from sklearn.datasets import load_iris
    iris=load_iris()
    # print(iris)
    # print(len(iris["data"]))#150个数据
    from sklearn.model_selection import train_test_split
    #把数据分为测试数据和验证数据 test_size=0.2验证数据集占20% 也就是150个数据有30个验证集 random_state=1随机的选择30个数据
    train_data,test_data,train_target,test_target=train_test_split(iris.data,iris.target,test_size=0.2,random_state=1)
    
    #2.进行建模
    from sklearn import tree
    #引入分类器
    clf=tree.DecisionTreeClassifier(criterion="entropy")
    clf.fit(train_data,train_target)#用训练集进行训练，一般情况下fit就是建立模型，而predict就是进行预测
    #进行预测
    y_pred=clf.predict(test_data)
    print(test_target)
    print(y_pred)
    import numpy as np
    print(np.mean(y_pred==test_target))

输出：
    [0 1 1 0 2 1 2 0 0 2 1 0 2 1 1 0 1 1 0 0 1 1 1 0 2 1 0 0 1 2]
    [0 1 1 0 2 1 2 0 0 2 1 0 2 1 1 0 1 1 0 0 1 1 2 0 2 1 0 0 1 2]
    0.9666666666666667

下面关于调参的内容是引用的别人的博客，原文网址：https://blog.csdn.net/qq\_38923076/article/details/82931340

sklearn中决策树算法参数共有13个，如下：

class sklearn.tree.DecisionTreeClassifier(criterion=’gini’, splitter=’best’, max\_depth=None, min\_samples\_split=2, min\_samples\_leaf=1, min\_weight\_fraction\_leaf=0.0, max\_features=None, random\_state=None, max\_leaf\_nodes=None, min\_impurity\_decrease=0.0, min\_impurity\_split=None, class\_weight=None, presort=False)  
英文文档在这：[决策树参数][Link 1]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQxMDc2Nzk3_size_16_color_FFFFFF_t_70][]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQxMDc2Nzk3_size_16_color_FFFFFF_t_70 1][]

# 回归： #

# 1.决策树回归
    from sklearn import tree
    
    model_decision_tree_regression = tree.DecisionTreeRegressor()

下面的参数说明转载出处为[https://blog.csdn.net/wl2858623940/article/details/80448875][https_blog.csdn.net_wl2858623940_article_details_80448875]

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQxMDc2Nzk3_size_16_color_FFFFFF_t_70 2][]

[Link 1]: https://blog.csdn.net/qq_38923076/article/details/82931340
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQxMDc2Nzk3_size_16_color_FFFFFF_t_70]: https://image.dandelioncloud.cn/pgy_files/images/2024/04/18/45efb036c6914cc9a97a29be1361087b.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQxMDc2Nzk3_size_16_color_FFFFFF_t_70 1]: https://image.dandelioncloud.cn/pgy_files/images/2024/04/18/647f4590cc344a0ca950a13dfca979b9.png
[https_blog.csdn.net_wl2858623940_article_details_80448875]: https://blog.csdn.net/wl2858623940/article/details/80448875
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQxMDc2Nzk3_size_16_color_FFFFFF_t_70 2]: https://image.dandelioncloud.cn/pgy_files/images/2024/04/18/1577e73269244fcc82eae53e8fec99b2.png