【摘要】 针对动态网络相邻网络快照社团检测结果差异较大、准确性不足的问题,文章在基于角色的标签扩散算法(role-based label propagation algorithm, RLPA)基础上,结合前一时刻社团结构信息以提高时序平滑性,根据时序结构差异性调整前序信息对当前社团结构的贡献权重,提出了一种基于角色的演化标签扩散算法(role-based evolutionary label propagation algorithm, RELPA)。在两类真实数据集上进行了对比实验,结果表明,该文算法能够准确有效地检测出动态网络中的社团结构。更多还原

【Abstract】 For the problems that existing static community detection algorithms cannot derive the results with temporal smoothness and may lead to poor results in dynamic networks, role-based evolutionary label propagation algorithm(RELPA) was proposed. Based on role-based label propagation algorithm(RLPA), the algorithm combined previous community structure information to improve the temporal smoothness and introduced evolutionary structural differences information as the weight of contribution of previous local community structure information to improve the accuracy. Time complexity analysis indicates that the proposed algorithm inherits the prominent speed of label propagation algorithm. The proposed algorithm was evaluated on two real networks. Results of experiments show that it can accurately detect evolutionary communities in dynamic networks.更多还原