Abstract: As to short message text clustering, this paper designs a hybrid clustering algorithm combining by frequent termsets and Ant-Tree algorithm. This algorithm takes the advantage of efficiency of processing text data based on the frequent termsets clustering, produces the initial cluster, then eliminates the overlap text documents by calculating silhouette coefficient. Further refines the cluster by Ant-Tree. Thus gets the high quality clustering results. And the results that retain the description and tree structure can provide wider applications.

Key words: frequent term-sets, Ant-Tree algorithm, silhouette coefficient, short message, clustering