基于禁忌搜索的聚类簇数目估算算法

An Estimation Algorithm for Number of Clusters Based on Tabu Search

本文提出一种基于禁忌搜索的聚类算法,能自动估计数据集的正确的聚类簇数目,其数据对象既可是紧凑球型分布,也可是非紧凑球型分布,或皆具两者。该算法包括三个阶段:最小生成树聚类、聚类簇重组和禁忌搜素优化。实验结果表明,所提出的算法可行、有效。

In this article, tabu search is used to automatically explore the proper number of clusters of data sets whose clusters are of either spherical or non-spherical shape. The proposed algorithm, TABU-Clustering, is com- posed of three stages: minimum spanning tree clustering, recombination of clusters and tabu search optimization. Computer simulations show that the proposed algorithm is feasible and effective.