点击次数：

论文类型：期刊论文

发表时间：2008-12-01

发表刊物：Journal of Computational Information Systems

收录刊物：EI、Scopus

卷号：4

期号：6

页面范围：2841-2847

ISSN号：15539105

摘要：In this paper, we present an agglomerative hierarchical algorithm based on the definition of cluster similarity (CS). Inspired by the node similarity of social networks, we give the definition of CS based on the common connecting strength. Because CS represents the similarity between the clusters by considering the inner and outer structure, the algorithm may obtain nonlocal results. Experiments on the public textual dataset indicate that the modularity, F-measure and entropy obtained by our algorithm are better than the one of FN algorithm and UPGMA. The compute complexity of out algorithm is O (mn ), and it may be suitable for analyzing large or dynamical data. © 2008 Binary Information Press.