Abstract: Community division is a hot topic in the study of bipartite network, aiming at the existing bipartite network community detecting algorithm from different nodes with the problem of low accuracy, this paper proposes the bipartite network community detecting algorithm based on Intimacy and Attraction Algorithm（IAA）. The algorithm treats every U-type node as a community, through calculating the intimacy of each community and the attraction of the community merge communities, U-type node partition is obtained. At last, V-type nodes are divided into existing communities to obtain complete community division results. Through the analysis on artificial networks and real-world networks, normalized mutual information and modularity are used respectively as evaluation indicators. The experimental results show that IAA is able to mine the bipartite network community structure more effectively and has a better community division.

Key words: bipartite network, community detecting, intimacy, attraction, normalized mutual information, modularity

摘要： 社区划分是二分网络研究中的一个热门话题，针对现有的二分网络社区发现算法存在从不同节点出发社区划分准确率低的问题，提出了基于亲密度和吸引力的二分网络社区发现算法（Intimacy and Attraction Algorithm，IAA）。该算法将[U]类中的每一个节点看作一个社区，通过计算出每一个社区的亲密度和社区间的吸引力来合并社区，从而得到[U]类节点的划分，最后[V]类节点划分到已有的社区中得到完整的社区划分结果。在人工数据集和真实网络上进行分析，分别利用互信息和模块度作为评价指标，实验结果表明，IAA能够更有效挖掘二分网络社区结构，具有良好的社区划分效果。

关键词: 二分网络, 社区发现, 亲密度, 吸引力, 归一化互信息, 模块度