I have a follow-up question related to sample size for cluster analysis.
Many recommend using 2^n - n is number of variables.
I would like to ask if we should adjust for finite population (n0/(1+n0/N) as we do with other methods of analysis?
Thank you.
Hao