In solving the clustering problem, traditional methods, for example, the K-
means algorithm and its variants, usually ask the user to provide the numbe
r of clusters. Unfortunately, the number of clusters in general is unknown
to the user. Therefore, clustering becomes a tedious trial-and-error work a
nd the clustering result is often not very promising especially when the nu
mber of clusters is large and not easy to guess. In this paper, we propose
a genetic algorithm for the clustering problem. This algorithm is suitable
for clustering the data with compact spherical clusters. It can be used in
two ways. One is the user-controlled clustering, where the user may control
the result of clustering by varying the values of the parameter, w. A smal
l value of w results in a larger number of compact clusters, while a large
value of w results in a smaller number of looser clusters. The other is an
automatic clustering, where a heuristic strategy is applied to find a good
clustering. Experimental results are given to illustrate the effectiveness
of this genetic clustering algorithm. (C) 2000 Pattern Recognition Society.
Published by Elsevier Science Ltd. All rights reserved.