任务7.Titanic-模型融合

1.投票法

from sklearn.ensemble import VotingClassifier
dtc=DecisionTreeClassifier()
svc = SVC(probability=True)
rfc = RandomForestClassifier(n_estimators=10, criterion="entropy")
lr = LogisticRegression() 
clf_vc = VotingClassifier(estimators=[('lr',lr),('dtc',dtc),('rfc',rfc),('svc',svc)])
clf_vc.fit(x_train, y_train)
print(clf_vc.score(x_train, y_train))

输出：0.9117174959871589

2.stacking融合

from mlxtend.classifier import StackingCVClassifier
sclf = StackingCVClassifier(classifiers=[dtc, rfc,svc],meta_classifier=lr,use_probas=True)
sclf.fit(x_train.values,y_train.values)
print(sclf.score(x_train, y_train))

输出：0.9454253611556982

Stacking融合方法的结果更优。

参考：https://blog.csdn.net/a5139515/article/details/79714111
https://blog.csdn.net/E22644/article/details/85833892