0

I am looking into Factor Analysis as a tool for dimensional reduction in a data set I am working with. I have a quick question regarding this technique.

How can I check that FA with a given number of common factors n captures enough information from my original data? I am looking for a criterion to assess variance retained when I convert my features similar to what we can do on PCA when we look at the eigenvalues after svd decomposition.

Thanks a lot for your reply.

Lumbreras
  • 51
  • 6
  • Perhaps some of the thoughts in http://stats.stackexchange.com/questions/236229/using-factor-analysis-as-a-data-reduction-tool?rq=1 may help. Be sure to read all the comments which contain useful nuggets. – mdewey Dec 07 '16 at 16:48
  • Hi! Thanks a lot but the post you refer to includes a suggestion on what seems to me filter feature selection methods and how to extract the scores. I am looking for a way to assess my scores capture the diversity in the data correctly. – Lumbreras Dec 09 '16 at 13:04

0 Answers0