Barclays Interview Questions | K mean
Question
How can we make sure that K-Means output is not sensitive to initialization?
in progress
1
Statistics
55 years
1 Answer
747 views
Great Grand Master 0
Answer ( 1 )
K-Means is a relatively efficient method.
However, specifying the number of clusters, in advance therfore the final results are sensitive to initialization but it often terminates at a local optimum.
I haven’t come across any other method to make it less sensitive during my career. Other people can please add