sklearn学习（2018.6.13）

学习使用sklearn模块，完成题目

题目如下：

代码实现：

from sklearn import datasets  
from sklearn import cross_validation  
from sklearn.naive_bayes import GaussianNB  
from sklearn.svm import SVC  
from sklearn.ensemble import RandomForestClassifier  
from sklearn import metrics  
  
#step1
dataset = datasets.make_classification(n_samples = 1500,n_features = 15);  
  
#step2 
k_fold = cross_validation.KFold(n = 1500 ,n_folds = 10,shuffle=True)  
for train_index,test_index in k_fold:  
    x_train,y_train = dataset[0][train_index],dataset[1][train_index]  
    x_test,y_test = dataset[0][test_index],dataset[1][test_index]  
      
#step4 
def evaluate(y_test,perd,name):  
    print(name+': ')  
    acc = metrics.accuracy_score(y_test,pred)  
    print('Accuracy: ',acc)  
    f1 = metrics.f1_score(y_test,pred)  
    print('F1_score: ',f1)  
    auc = metrics.roc_auc_score(y_test,pred)  
    print('AUC ROC: ',auc)  
  
#SVC      
C_values = [1e-02, 1e-01, 1e00, 1e01, 1e02]  
for C_value in C_values:  
    clf = SVC(C=C_value, kernel='rbf', gamma=0.1)  
    clf.fit(x_train, y_train)  
    pred = clf.predict(x_test)  
    evaluate(y_test,pred,"SVC, C_value= %s" % str(C_value))  
  
#RandomForestClassifier  
clf = RandomForestClassifier(n_estimators = 6)  
clf.fit(x_train,y_train)  
pred = clf.predict(x_test)  

evaluate(y_test,pred,'RandomForestClassifier')  

#step3
  
# GaussianNB  
clf = GaussianNB()  
clf.fit(x_train,y_train)  
pred = clf.predict(x_test)  
evaluate(y_test,pred,'GaussianNB')