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。
The proposed algorithm, TABU-Clustering, is com - posed of three stages: minimum spanning tree clustering, recombination of clusters and tabu search optimization.
然后使用最小生成树方法得到聚类的初始结构；
The initial clustering structure is constructed by using the minimum spanning tree method.
最后，给出了改进的基于划分的聚类方法，以及其中涉及到的最小生成树算法的基本思想。
Last, the partition algorithm which has been improved is proposed, meanwhile, the basic idea of Minimum Cost Spanning tree is presented.
应用推荐