similarity measure

nguyenxuanhau

Hi
In K-mean algorithm of Rapid, what was the measure used?
Thanks

pep

The distance measure (which is opposed to the similarity measure) in k-means should be the Euclidian distance, by default.

nguyenxuanhau

Hi
So, can we change the similarity measure in K-mean algorithm to Cosine measure? because many articles evaluated that Cosine or jaccard measures is better than Euclidean measures
best regard
Thanks