# How can I select features for a symbolic regression problem to be solved with genetic programming?

I want to solve a symbolic regression problem with genetic programming. My dataset is similar to this one, but I have 30 features, and I want to use only the most sensitive features. I found this library interesting for Symbolic Regression, but could not find the right approach for feature selection.

• Have you tried to look at the existing feature selection techniques typically used in machine learning? If yes, why aren't they good for you? If not, it may be a good idea to have a look at them. Once you know the answer to your question, feel free to write it below (if it's good ehough) ;) – nbro Mar 7 at 11:20
• You should probably read this: Feature Selection to Improve Generalization of Genetic Programming for High-Dimensional Symbolic Regression. – nbro Mar 10 at 15:50
• I do not have full access to that paper, but it claims as follows, In this paper, we propose a new feature selection method based on permutation to select features for high-dimensional SR using GP. So, permutation using GP. what's it really? – thunder Mar 10 at 16:46
• You can find a freely available version of this paper (or a draft) on the web. – nbro Mar 10 at 16:47
• Is there a python library for GP? – thunder Mar 10 at 17:00