dgp, I am not qualified to answer any of your questions, but I thought you might want to look at the link which I believe is a version of kmeans cluster analysis in J. I have some experience with hierarchical cluster analysis, but not with kmeans. The code at this link has not been widely tested nor evaluated for efficiency. But there are many clever J coders on this list which who I predict would we helpful in refining such an algorithm.
http://www.jsoftware.com/jwiki/BrianSchott/code/nonhier On Tue, Aug 4, 2009 at 11:04 AM, Plugge NIH<[email protected]> wrote: > J Users, > [snip] > > One CPU intensive routine we employ is Cluster Analysis using a K-Means > algorithm to find cluster centroids. > [snip] > > Sincerely, > dgp -- (B=) <-----my sig Brian Schott ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
