I have some features (predictor variables) that I want to regress on a response. The predictor features are correlated with each other. Hence, for performing regression I need to do a principal component analysis first, compute new uncorrelated variables and then perform linear regression.

I also calculated partial correlations of each predictor variable with response by controlling for some reported confounding factors.

My question is - do variables showing high partial correlations with response, also explain a high proportion of variance explained by a principal component? Is it necessary? If not how to compare PCA and partial correlation results? Which is better? I'm performing PCA-based regression for the first time and am a complete novice. Please help.


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.