我有:
r =包含6个特征的numpy数组,即形状〜10000 x 6
我想学习预测的answer =二进制numpy数组(1或0),形状为〜1000 x 1
我已经完成:
[datatrain, datatest, answertrain, answertest] = cross_validation.train_test_split(r,answer)
clf = linear_model.ElasticNet()
clf.fit(X=datatrain,y=answertrain)
prediction = clf.predict(datatest)
如何获得ROC的阈值?我看不到决策函数或预测概率的属性。