Interpreting Principal Component Regression
Posted: Tue May 12, 2020 4:27 pm
I'm looking to resolve multicollinearity by using PCA. So far, I've managed to create the principal components I need, but I'm trying to find a way in views to combine these into interpretable coefficients in a regression on my dependent variable. Ideally, I'd like to use the components to create new coefficients on the original explanatory variables, but I'm not sure how to go about doing this in Eviews. Is this possible and if so, how can I go about it?