File history uploaded by mike@mbowles.com 12 years, 8 months ago

In class, we constructed a clustering problem by generating gaussian clusters of 100 points around each of 25 centers - centers with coordinates (i,j) i = 1 to 5 and j = 1 to 5.  See what number of clusters looks like the best clustering of this system for for each of three situations sd = 0.1, sd = 0.2 and sd = 0.5.  Use both the inflection point technique and the AIC (Akaike information criterion) to make your determination.

Comments (0)

You don't have permission to comment on this page.