In Python sklearn.svm.OneClassSvm, how to calculate ROC and AUC only with predicted label and real label?

Question

I am using Python 3.6,sklearn.svm.OneClassSVM to practice OSVM and I want to

calculate ROC, AUC.

I have used decision_function() to calculate ROC and AUC ,the code is below.

I want to evaluate the value that I calculate by decision_function.

Can I only use predicted label and real label to obtain ROC, AUC value?

     y_score = oneclass.decision_function(testing_data)
     roc_auc = metrics.roc_auc_score(Y_test, y_score)

score 0 · Answer 1 · answered May 20 '19 at 08:56

0

I am not sure if I get your question complete correctly, but if you do this:

clf = svm.OneClassSVM(nu=0.1, kernel="rbf", gamma=0.1)
clf.fit(X_train)
y_pred_train = clf.predict(X_train)
y_score = clf.predict(X_test)

Then you should be able to use:

from sklearn.metrics import roc_auc_score
roc_auc_score(y_test, y_score)
``

answered May 20 '19 at 08:56

PV8

Do you mean that y_score should be 1 or -1? Thanks for answer my question. – Tina May 21 '19 at 05:47
it depends what is your target value, but in generel this should work – PV8 May 22 '19 at 10:51

1 Answers1