Dear researchers,
I have a positive class that I trained use one-class-classification (libsvm-library). How can I test (svmpredict) using negative class and then using positive class separately or testing the one-class is true or not?
Thanks
I believe you have to train your classifier for both - positive and negative examples first, then you can evaluate the performance separately on positive and then negative examples only. Never the less I would evaluate the test performance both pos & neg examples in a single set.
you can use weka tool for one-class classification support vector machine
here is a tutorial:
https://weka.wikispaces.com/How+do+I+perform+one-class+classification%3F?responseToken=0fd40e53c8908b3aab5d7aa2a84951641
Train your classifier by using both cases (+ and -) first. Then evaluate whether the classifier is able to identify the class labels correctly.
- Badges
- Science topic
- Similar topics
- Organizational Psychology
- Training