0

Is it correct to only rescale a large int feature and leave the % features as is when running k-means algorithm for segmentation analysis ?

e.g. a feature is the population of the city(150 000), then the rest features are % (e.g. 0.46) of energy consumption etc etc.

Proddie
  • 1
  • possible duplicate of https://stats.stackexchange.com/questions/21222/are-mean-normalization-and-feature-scaling-needed-for-k-means-clustering ? – Regressor Apr 12 '22 at 20:36
  • I've actually read that thread prior to posting this.to my mind the difference of this post is whether it's appropriate to only rescale one of them rather than both continuous variables/features. – Proddie Apr 13 '22 at 21:05
  • I think the best answer is to `test` and `evaluate` with the change and without the change. – Regressor Apr 13 '22 at 21:33

0 Answers0