中文摘要：

结合复杂网络社团结构的相关研究,提出一种基于网络社团结构和模块化函数的聚类算法CSMFBCA（Community Structure and Modularization Function Based Clustering Algorithms）。算法通过数据点之间的关系进行融合,形成一定的数据簇,然后定义一个统筹全局的模块化函数,再通过最大化模块函数值,得到最优的聚类结果。实验结果表明,该算法不仅能很好地解决凹形数据聚类以及聚类个数识别的问题,而且能处理权重无向网络的社团发现问题,比现有的典型算法有明显的优势。

英文摘要：

In conjunction with related study on complex network community structure,in this paper we propose a clustering algorithm CSMFBCA,which is based on network community structure and modular function. The algorithm forms certain data cluster by fusing the data through the relationship among data points,and then defines a modular function which co-ordinates the whole,finally the optimal clustering results are obtained through maximising the modular function value. Experimental results show that this algorithm can gracefully cope with concave data clustering and the identification of clusters number. Furthermore,it can also deal with community detection issue in weighted and undirected network,thus has remarkable advantages over current typical algorithms.