提出了一种直接基于原始二分网的聚类方法,该方法利用二分网上的资源分配过程得到资源分布矩阵,然后用模糊聚类方法对向量进行聚类,用F统计量判定最优的聚类结果.该方法在探测社团结构明显的人工网时具有较高的准确度.
A clustering algorithm on original bipartite network is proposed. Through resource allocation on bipartite network, a distribution matrix of resources is obtained. Fuzzy clustering method and F statistic are used to give the best community structure. This algorithm has high accuracy when detecting community structure of artificial bipartite networks which have obvious structure.