2
$\begingroup$

I'm going to implement a feature selection algorithm, and I plan to use the F-score for because of its simplicity. The problem is that, the F-score is used for binary classification. How can it be extended for multi-class classification?

My idea is to use one-vs-all method. That is, if I have k classes, I would generate k F-scores for each of these (by taking only one feature at a time and calculate the F-score against all others) and then just take the average. Would this be a good approach?

$\endgroup$
1
$\begingroup$

F-score tends to be very similar to likes of t and Kruskal-Wallis tests, when it comes to feature ranking. So, one solution can be using multi-level alternatives of t-test, e.g., ANOVA or their non-parametric versions.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.