Currently trying to use Monte-Carlo Tree Search to automate the process of feature selection for SVM (which I'm using to evaluate my features). Although successful so far, It tightly depends on the speed of the evaluation function (SVM) which becomes impractically slow for datasets with a large number of training examples. I just wanted to have your opinion on this approach and whether there are other methods that could be useful to evaluate features without having to run a model on them.

  • $\begingroup$ boruta does feature selection using random forests which might or might not be speedier than fitting a bunch of SVMs $\endgroup$ – Sycorax Apr 9 at 20:36

Your Answer

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

Browse other questions tagged or ask your own question.