A Data Science Central Community
I have been using Proc Varclus for dimensionality reduction in SAS. It clusters variables and gives you the 1-R**2 ratio to pick up the most representative variable from the cluster and leave out the rest. I have not used Princomp as it is difficult to preserve original variables with it and we need them to interpret the business meaning of the model. Is there a way I can do that with princomp or SVD? Please share any examples as well.