Should feature scaling be used while using unsupervised algorithms?

Question

I have read many articles and resources about using feature scaling and when to use it, in particular two answers on this website as well-

When should I apply feature scaling for my data?

What algorithms need feature scaling, beside from SVM?

But all answers focus on using feature scaling on supervised algorithms.

My question is - Should I use feature scaling in unsupervised algorithms (like clustering) as well? Or, feature scaling makes no sense in unsupervised problems? Or, feature scaling has no effect in unsupervised problems?

Answer 1

Yes, in most cases you should also scale your data. For instance k-means is scaling dependent as it computes a distance between the two samples. PCA is also scaling dependent, as most of the variance (in absolute term) in your data would occur for the variable that has the largest scale/values.