I only used the cases (150 subjects) and run the cluster analysis using FASTCLUS procedure and end up with three cluster solution. Then I plotted the cluster grouping using the canonical variables through canonical discriminant analysis. I didn't used the matched controls patients in the cluster analysis. Now, I need to use the matched controls (which has only 35 pateints) along with 35 cases. The total subjects now reduced to 70 only. Can I use the same FASTCLUS procedure to identify the cluster? OR Is there any other sas procudure available to identify the cluster solution for the matched case control cohort? Thanks GKB
... View more