News for hackers. Stuff that matters.
Re:Machine Learning: Curse of Dimensionality by NhanH on Wednesday August 05, 2015 @02:03PM (#10011597)
You're supposed to normalize the value of your range on all dimensions so that they are the same numerical magnitude/range (as the simplest normalization).

The issue is definitely not specific to clustering, you would have the same problem with one dimension dominate the weight without normalization even for simple regression too.

There are no comments on this comment yet.