Hi all.
I am using PCA to generate PCs and then using them as independent variables in regression.
I added an extra data-step after PCA which is transformation of the PCs e.g. X=prin1^2
This significantly improved the subsequent regression model, however I am not sure if transforming the PCs is not a common/suggested practice for reasons that I am not aware of...
Thank you!
I am using PCA to generate PCs and then using them as independent variables in regression.
I added an extra data-step after PCA which is transformation of the PCs e.g. X=prin1^2
This significantly improved the subsequent regression model, however I am not sure if transforming the PCs is not a common/suggested practice for reasons that I am not aware of...
Thank you!