A genetic clustering algorithm for data with non-spherical-shape clusters

In solving clustering problem, traditional methods, for example, the K-means algorithm and its variants, usually ask the user to provide the number of clusters. Unfortunately, the number of clusters in general is unknown to the user. The traditional neighborhood clustering algorithm usually needs the user to provide a distance d for the clustering. This d is difficult to decide because some clusters may be compact but others may be loose. In this paper, we propose a genetic clustering algorithm for clustering the data whose clusters are not of spherical shape. It can automatically cluster the data according to the similarities and automatically find the proper number of clusters. The experimental results are given to illustrate the effectiveness of the genetic algorithm. (C) 2000 Pattern Recognition Society. Published by Elsevier Science Ltd. All rights reserved.