week15

Assignment

from sklearn import datasets
from sklearn.naive_bayes import GaussianNB
from sklearn import cross_validation
from sklearn.svm import SVC
from sklearn import metrics

# 生成数据集
data = datasets.make_classification(n_samples=1000, n_features=10)
test = 0;
X, Y = data[0], data[1];
kf = cross_validation.KFold(len(Y), n_folds=10, shuffle=True)
for train_index, test_index in kf:
    X_train, y_train = X[train_index], Y[train_index]
    X_test, y_test = X[test_index], Y[test_index]
    test += 1;
    print("\nTEST ", test);
#朴素贝叶斯Naive Bayes
    clf = GaussianNB()
    clf.fit(X_train, y_train)
    pred = clf.predict(X_test)
    print("Naive Bayes: ");
    acc = metrics.accuracy_score(y_test, pred)
    print ("acc: ", acc)
    f1 = metrics.f1_score(y_test, pred)
    print ("f1: ", f1)
    auc = metrics.roc_auc_score(y_test, pred)
    print ("auc: ", auc)


#SVM
    clf = SVC(C=1e-01, kernel='rbf', gamma=0.1)
    clf.fit(X_train, y_train)
    pred = clf.predict(X_test)
    print("\nSVM: ");
    acc = metrics.accuracy_score(y_test, pred)
    print ("acc: ", acc)
    f1 = metrics.f1_score(y_test, pred)
    print ("f1: ", f1)
    auc = metrics.roc_auc_score(y_test, pred)
    print ("auc: ", auc)

#Random Forest
    from sklearn.ensemble import RandomForestClassifier
    clf = RandomForestClassifier(n_estimators=6)
    clf.fit(X_train, y_train)
    pred = clf.predict(X_test)
    print("\nRandom Forest: ");
    acc = metrics.accuracy_score(y_test, pred)
    print ("acc: ", acc)
    f1 = metrics.f1_score(y_test, pred)
    print ("f1: ", f1)
    auc = metrics.roc_auc_score(y_test, pred)
    print ("auc: ", auc)

TEST  1
Naive Bayes: 
acc:  0.84
f1:  0.8431372549019607
auc:  0.8399359743897561

SVM: 
acc:  0.85
f1:  0.8484848484848485
auc:  0.8505402160864346

Random Forest: 
acc:  0.9
f1:  0.9019607843137255
auc:  0.8999599839935973

TEST  2
Naive Bayes: 
acc:  0.91
f1:  0.9072164948453608
auc:  0.9134460547504026

SVM: 
acc:  0.93
f1:  0.924731182795699
auc:  0.9303542673107892

Random Forest: 
acc:  0.91
f1:  0.9052631578947369
auc:  0.9118357487922706

TEST  3
Naive Bayes: 
acc:  0.92
f1:  0.923076923076923
auc:  0.920915295062224

SVM: 
acc:  0.92
f1:  0.9199999999999999
auc:  0.9233239662786029

Random Forest: 
acc:  0.95
f1:  0.9523809523809524
auc:  0.9504215174628663

TEST  4
Naive Bayes: 
acc:  0.87
f1:  0.8571428571428571
auc:  0.8677238057005219

SVM: 
acc:  0.87
f1:  0.853932584269663
auc:  0.8665194700923324

Random Forest: 
acc:  0.93
f1:  0.924731182795699
auc:  0.9291449217181854

TEST  5
Naive Bayes: 
acc:  0.87
f1:  0.8631578947368421
auc:  0.8693910256410255

SVM: 
acc:  0.92
f1:  0.9130434782608695
auc:  0.9182692307692308

Random Forest: 
acc:  0.95
f1:  0.946236559139785
auc:  0.9487179487179486

TEST  6
Naive Bayes: 
acc:  0.94
f1:  0.9473684210526316
auc:  0.95

SVM: 
acc:  0.89
f1:  0.8990825688073394
auc:  0.9083333333333333

Random Forest: 
acc:  0.88
f1:  0.890909090909091
auc:  0.8958333333333334

TEST  7
Naive Bayes: 
acc:  0.91
f1:  0.8915662650602411
auc:  0.9088466308391897

SVM: 
acc:  0.91
f1:  0.8860759493670887
auc:  0.9014055394791236

Random Forest: 
acc:  0.95
f1:  0.9382716049382716
auc:  0.9464654816039686

TEST  8
Naive Bayes: 
acc:  0.89
f1:  0.897196261682243
auc:  0.8944805194805197

SVM: 
acc:  0.91
f1:  0.9158878504672897
auc:  0.9147727272727273

Random Forest: 
acc:  0.91
f1:  0.9174311926605505
auc:  0.9123376623376623

TEST  9
Naive Bayes: 
acc:  0.87
f1:  0.8686868686868686
auc:  0.8701480592236894

SVM: 
acc:  0.88
f1:  0.8775510204081631
auc:  0.8799519807923168

Random Forest: 
acc:  0.91
f1:  0.9010989010989012
auc:  0.908563425370148

TEST  10
Naive Bayes: 
acc:  0.9
f1:  0.8979591836734694
auc:  0.8999599839935974

SVM: 
acc:  0.9
f1:  0.8936170212765957
auc:  0.8991596638655462

Random Forest: 
acc:  0.91
f1:  0.9072164948453607
auc:  0.909763905562225

对比数据可以看出，Random Forest算法的结果要好。