I have two different set of features for which I have a SVM classifier and a Naive Bayes classifier, respectively. If I wanted to combine these two classifiers to get a better prediction, what option do I have?
- Anybody can ask a question
- Anybody can answer
- The best answers are voted up and rise to the top
You should look into the variety of ensemble methods available for this sort of situation! One idea you might try, is try classifying every sample with both classifiers, and then take the vote of the classifier that has the most certainty associated with its prediction. I like this particular approach, because I tend to learn something new about the way I'm modeling my features by looking at the samples that both classifiers get wrong.