0

If the answer is no, could someone give a simple counterexample?

Thank you!

RR_28023
  • 95
  • 4
  • 1
    Could you clarify the interplay between PCA and k-means in your setup? Are you running PCA then k-means? – chl Nov 07 '20 at 13:03
  • Hi. I am running a k-means clustering algorithm using 20 variables. Separately I am using PCA analysis to obtain the first two principal components, and checking which of my variables contributed the most to those principal components. Asking if those variables are also the ones that have more weight in the k-means clustering process. Thanks – RR_28023 Nov 09 '20 at 14:07
  • How do you determine variable weight in k-means? – chl Nov 09 '20 at 16:13
  • That is exactly what I want to know (how to estimate the weight of each variable in the k-means). I am asking if checking the variables that contribute to the top Principal Components is a valid approximation. – RR_28023 Nov 10 '20 at 17:13
  • Ah, I see now. You will probably find useful information in this [related thread](https://stats.stackexchange.com/q/183236/930). If this is the case, then I believe this thread should be closed as a duplicate. Otherwise, please clarify. – chl Nov 11 '20 at 06:21
  • Thank you! Indeed that thread addresses my question – RR_28023 Nov 11 '20 at 16:15

0 Answers0