map - K-means algorithm variation with equal cluster size -


I find the fastest algorithm for the group based on distance in groups of sizes evenly on points on a map I am here. Looks straight and promising, but does not produce evenly shaped groups.

Is there any difference in this algorithm or is there a different one which allows the same number of members for all groups?

See also:

Can be: Emonic centrosides sort the centrosides by descending in the shape of their linked groups in an array. For k -1 through i = 1, the data points in the cluster i are any other centroid j ( I & lt; j รข ?? ¤ k ) j for closed and centroid recompute I (but do not compress the cluster again) unless the cluster size is n / k .

The complexity phase of this post process is ( k one class n LG n ).

Comments