特征选择--PCA与互信息的使用与比较

Handsome_zhou

1. from sklearn.datasets import load_boston
   
2. import pandas as pd
   
3. import numpy as np
   
4. d = load_boston()
   
5. x = d.data
   
6. y = d.target
   
7. print(x[:10])
   
8. print('形状:', x.shape)
   
9. 
   
10. from sklearn.decomposition import PCA
    
11. pca = PCA(n_components=10)
    
12. x = pca.fit_transform(x)
    
13. 
    
14. %time
    
15. x = pd.DataFrame(d.data)
    
16. y = pd.DataFrame(d.target)
    
17. x = np.matrix(x)
    
18. y = np.matrix(y)
    
19. 
    
20. from sklearn.preprocessing import StandardScaler #标准化
    
21. scaler = StandardScaler()
    
22. X = scaler.fit_transform(X)
    
23. y = scaler.fit_transform(y)
    
24. 
    
25. from sklearn.model_selection import train_test_split
    
26. X_train,X_test,y_train,y_test = train_test_split(X,y,test_size = 0.2)
    
27. 
    
28. from sklearn.svm import SVR
    
29. svr = SVR(gamma = 'scale')
    
30. 
    
31. svr.fit(X_train,y_train)
    
32. train_score = svr.score(X_train,y_train)
    
33. test_score = svr.score(X_test,y_test)
    
34. print('train score: {} ; test score: {}'.format(train_score,test_score))

复制代码

结果：

1. from sklearn import datasets
   
2. from sklearn.feature_selection import SelectKBest
   
3. from sklearn.feature_selection import mutual_info_regression
   
4. import pandas as pd
   
5. import numpy as np
   
6. from sklearn.datasets import load_boston
   
7. d = load_boston()
   
8. 
   
9. x = pd.DataFrame(d.data)
   
10. y = pd.DataFrame(d.target)
    
11. x = np.matrix(x)
    
12. y = np.matrix(y)
    
13. mi = mutual_info_regression(x,y)
    
14. X = SelectKBest(mutual_info_regression, k=10).fit_transform(x,y)
    
15. 
    
16. from sklearn.preprocessing import StandardScaler #标准化
    
17. scaler = StandardScaler()
    
18. X = scaler.fit_transform(X)
    
19. y = scaler.fit_transform(y)
    
20. 
    
21. from sklearn.model_selection import train_test_split
    
22. X_train,X_test,y_train,y_test = train_test_split(X,y,test_size = 0.2)
    
23. 
    
24. from sklearn.svm import SVR
    
25. svr = SVR(gamma = 'scale')
    
26. 
    
27. svr.fit(X_train,y_train)
    
28. train_score = svr.score(X_train,y_train)
    
29. test_score = svr.score(X_test,y_test)
    
30. print('train score: {} ; test score: {}'.format(train_score,test_score))

复制代码

结果：


在用boston数据集得到的结果来看，PCA和互信息两种特征选择方法得到的数据在支持向量机下并没有明显的性能差别。

tomok

学习