ML之DT：利用DT(DTC)实现对iris(鸢尾花)数据集进行分类并可视化DT结构

ML之DT：利用DT(DTC)实现对iris(鸢尾花)数据集进行分类并可视化DT结构

目录

输出结果

实现代码

输出结果

实现代码

  

   

    

     

    
    

     

      #1、
     
    
   

    

     

    
    

     

      iris = load_iris()
     
    
   

    

     

    
    

     

      dir(iris)
     
    
   

    

     

    
    

     
 
     
    
   

    

     

    
    

     
 
     
    
   

    

     

    
    

     

      iris_feature_name = iris.feature_names
     
    
   

    

     

    
    

     

      iris_features = iris.data
     
    
   

    

     

    
    

     

      iris_target_name = iris.target_names
     
    
   

    

     

    
    

     

      iris_target = iris.target
     
    
   

    

     

    
    

     
 
     
    
   

    

     

    
    

     

      print('iris_feature_name','\n',iris_feature_name)
     
    
   

    

     

    
    

     

      print('iris_features前5','\n',iris_features[:5,:],iris_features.shape)
     
    
   

    

     

    
    

     

      print('iris_target_name','\n',iris_target_name)
     
    
   

    

     

    
    

     

      print('iris_target','\n',iris_target)
     
    
   

    

     

    
    

     
 
     
    
   

    

     

    
    

     
 
     
    
   

    

     

    
    

     

      #2、
     
    
   

    

     

    
    

     

      clf = tree.DecisionTreeClassifier(max_depth=4)
     
    
   

    

     

    
    

     

      clf = clf.fit(iris_features, iris_target)
     
    
   

    

     

    
    

     
 
     
    
   

    

     

    
    

     
 
     
    
   

    

     

    
    

     

      #3、
     
    
   

    

     

    
    

     

      import pydotplus
     
    
   

    

     

    
    

     

      from IPython.display import Image, display
     
    
   

    

     

    
    

     
 
     
    
   

    

     

    
    

     
 
     
    
   

    

     

    
    

     

      dot_data = tree.export_graphviz(clf,
     
    
   

    

     

    
    

     

       out_file = None,
     
    
   

    

     

    
    

     

       feature_names = iris_feature_name,
     
    
   

    

     

    
    

     

       class_names = iris_target_name,
     
    
   

    

     

    
    

     

       filled=True,
     
    
   

    

     

    
    

     

       rounded=True
     
    
   

    

     

    
    

     

       )
     
    
   

    

     

    
    

     
 
     
    
   

    

     

    
    

     

      from IPython.display import display, Image
     
    
   

    

     

    
    

     

      graph = pydotplus.graph_from_dot_data(dot_data)
     
    
   

    

     

    
    

     

      # graph.write_png(r"DT.png")
     
    
   

    

     

    
    

     

      display(Image(graph.create_png()))
     
    
   

    

     

    
    

     

      Image(graph.create_png())
     
    
   

    

     

    
    

     
 
     
    
   

    

     

    
    

     

      import matplotlib.pyplot as plt
     
    
   

    

     

    
    

     
 
     
    
   

    

     

    
    

     

      img_path='DT.png'
     
    
   

    

     

    
    

     

      plt.imshow(img_path)
     
    
   

    

     

    
    

     

      plt.show()
     
    
  
 

文章来源: yunyaniu.blog.csdn.net，作者：一个处女座的程序猿，版权归原作者所有，如需转载，请联系作者。

原文链接：yunyaniu.blog.csdn.net/article/details/86551069