基于社交信息传播群体识别的主题重叠性k类群过滤算法

Topic overlap k-group filtering algorithm based on social information dissemination group recognition

结合复杂网络理论和派系过滤算法的特点,提出了一种主题重叠的k类群过滤算法。该算法通过迭代回归计算和设计重叠矩阵,能够有效识别具有兴趣重叠特征的节点及多兴趣主题的圈子。仿真实验结果表明,与传统的谱平分算法和派系过滤算法相比,k类群过滤算法在量化个体对消息的兴趣程度以及识别单一或多主题消息传播群体方面表现更佳。

A topic overlapping k-group filtering algorithm is proposed by combining the characteristics of complex network theory and faction filtering algorithms.This algorithm can effectively identify nodes with overlapping interests and circles with multiple interest topics through iterative regression calculation and design of overlapping matrices.The simulation experiment results show that compared with traditional spectral equalization algorithms and faction filtering algorithms,the kgroup filtering algorithm performs better in quantifying individuals'interest in messages and identifying single or multiple topic message propagation groups.