Aiming at the efficient search of alliance and key enterprises in trust network, a trust network clustering model was established and an overlapping community discovery algorithm(CDNW)based on node weight was proposed. The weights of nodes were set according to the trust between enterprises in the trust network,and the initial communities were started with the nodes with the heaviest weights.The overlap coefficients were used to merge the communities with large overlapping degree.The combined community was detected by the correlation degree, and the nodes with the smaller correlation degree were removed from the community and re-allocated to the community.Test different datasets, the experimental results show that the algorithm has a good ability to divide the datasets whose community structure is not obvious and has certain stability.%针对信任网络中联盟和关键企业的高效寻找问题,建立信任网络聚类模型,提出一种基于节点权重的重叠社区发现算法CDNW(overlapping community discovery based on node weights).根据信任网络中企业间的信任度设定节点的权重,以权重最大的节点开始划分初始社区;采用重叠系数将重叠度大的社区合并,合并后的社区用关联度进行检测,将关联度过小的节点移出社区并为它重新分配社区.对不同的数据集进行测试,实验结果表明算法有较好的划分社区结构不明显的数据集的能力,具有一定的稳定性.
展开▼
