Abstract
Abstract A BASIC program, KOLCLUS, is given for a nonparametric significance test for the distinctness of a pair of clusters in Euclidean space, using the Kolmogorov-Smirnov statistic. The input data may be given in one of two forms: (1) the coordinates of the points in the full space; or (2) the distances of the points to the cluster centroids together with the intercentroid distance. The test consists of comparing the cumulative distribution of the projections of the points onto the intercentroid axis with hypothesized distributions that include a number of cluster distributions of interest. The test has less power than the parametric equivalent, but is highly conservative.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
