2.x ESL第二章习题 2.8

最新推荐文章于 2024-10-10 11:54:20 发布

weixin_30765577

最新推荐文章于 2024-10-10 11:54:20 发布

阅读量164

点赞数

文章标签：人工智能 python

原文链接：http://www.cnblogs.com/porco/p/4724940.html

版权

题目

代码

 
   
        import 
        pandas as pd 
       
 
        import 
        numpy as np 
       
 
        train_full 
        = 
        pd.read_csv( 
        '../zip.train' 
        ,sep 
        = 
        ' ' 
        ,engine 
        = 
        'c' 
        ,header 
        = 
        None 
        ).values[:, 
        0 
        : 
        - 
        1 
        ] 
       
 
        test_full 
        = 
        pd.read_csv( 
        '../zip.test' 
        ,sep 
        = 
        ' ' 
        ,engine 
        = 
        'c' 
        ,header 
        = 
        None 
        ).values 
       
 
            
       
 
        train  
        = 
        np.vstack((train_full[train_full[:, 
        0 
        ] 
        = 
        = 
        2 
        ],train_full[train_full[:, 
        0 
        ] 
        = 
        = 
        3 
        ])) 
       
 
        test  
        = 
        np.vstack((test_full[test_full[:, 
        0 
        ] 
        = 
        = 
        2 
        ],test_full[test_full[:, 
        0 
        ] 
        = 
        = 
        3 
        ])) 
       
 
            
       
 
        train_x 
        = 
        train[:, 
        1 
        :] 
       
 
        train_y 
        = 
        train[:, 
        0 
        ] 
       
 
        test_x 
        = 
        test[:, 
        1 
        :] 
       
 
        test_y 
        = 
        test[:, 
        0 
        ] 
       
 
            
       
 
        dc  
        = 
        [] 
       
 
        from 
        sklearn.linear_model  
        import 
        LinearRegression 
       
 
        lrcf  
        = 
        LinearRegression() 
       
 
        lrcf.fit(train_x, train_y) 
       
 
        dc.append(( 
        'linear regression' 
        ,lrcf)) 
       
 
            
       
 
        from 
        sklearn.neighbors  
        import 
        KNeighborsClassifier 
       
 
        for 
        i  
        in 
        [ 
        1 
        , 
        3 
        , 
        5 
        , 
        7 
        , 
        15 
        ]: 
       
 
             
        knn  
        = 
        KNeighborsClassifier(n_neighbors 
        = 
        i) 
       
 
             
        knn.fit(train_x,train_y) 
       
 
             
        dc.append(( 
        '%d-nearest neighbor' 
        % 
        (i),knn)) 
       
 
            
       
 
        def 
        acc(clf,x,y): 
       
 
             
        res  
        = 
        clf.predict(x) 
       
 
             
        if 
        type 
        (clf) 
        = 
        = 
        LinearRegression: 
       
 
                 
        res[res> 
        2.5 
        ] 
        = 
        3 
       
 
                 
        res[res< 
        2.5 
        ] 
        = 
        2 
       
 
             
        n  
        = 
        y.shape[ 
        0 
        ] 
       
 
             
        r  
        = 
        res[(res 
        = 
        = 
        y)].shape[ 
        0 
        ] 
       
 
             
        return 
        r 
        * 
        1.0 
        / 
        n 
       
 
            
       
 
        for 
        i  
        in 
        dc: 
       
 
             
        accr  
        = 
        acc(i[ 
        1 
        ],test_x,test_y) 
       
 
             
        print 
        '%s:  %.5f' 
        % 
        (i[ 
        0 
        ],accr) 
       
 
 