svm one classification accuracy without labeled data

hhamidy 0
Viewed 19

I am trying to detect anomaly data from dataset with OneClassSVM. I have not input labels and Input dataset haven't any label.

how I calculate accuracy and recall , ... . Because when we have output label's we can calculate from this method:

#Check the prediction performance
print(classification_report(y_test, costomized_prediction))

And we can compare real label by prediction label. But now when we don't have real label how we can calculate accuracy, recall, ...?

svm dataframe python
1 Answers

Accuracy. recall etc are dependent on the ground truth (or labels). So you need to hunt for them to understand how your model is performing.

the_ordinary_guy0
Related