There are two methods for feature selection with a particular classifier:

1. Exhaustive feature selection

2. Sequential forward selection

Exhaustive feature selection method performs more detailed evaluation the features in will give you more appropriate result but is computationally expensive while sequential forward selection is more faster than exhaustive feature selection

For dimensional reduction use PCA

Thanks sir, but i don't want to use the classifier (wrapper) so i used the filter approach.

is the PCA considered filter approach?

What features are you using?

Hi, please check out my most publication on my page where i used filter feature selection techniques. This article also contains some good references that will help you too. You can download the article from researchgate page. Hope that helps

Danielle