高级检索

    基于话题综合因子分析的语义社会网络社区发现算法

    Semantics Social Network Community Detection Algorithm Based on Topic Comprehensive Factor Analysis

    • 摘要: 针对一般社会网络社区发现算法仅考虑各节点的邻接关系,所划分的社区仅为一元关系社区,不能代表社区成员的语义相似性且无法处理具有多元语义话题的语义社会网络社区发现问题,提出基于话题因子分析的语义社会网络社区发现算法.该算法将节点的多元信息抽象为话题,先以多元话题综合因子作为节点话题信息度量,以节点间的话题密度差异作为节点聚合方向,构建初始社区结构;再以最大化社区内部话题信息相似度和最小化社区外部话题信息相似度为目标建立语义社区发现的目标函数及节点变动的代价函数;再以初始社区结构和代价函数作为初始解和判断准则,以节点变动的代价函数值为参数,建立全局优化的模拟退火策略优化语义社区结构,实现语义社会网络的语义社区发现;最后通过实验分析验证了算法的有效性.

       

      Abstract: Aiming at the problem that general social network community detection algorithm could only detect the community with sole relationship, which couldn't represent the semantics similarity of real social community and couldn't resolve the problem of SSN community detection with multiple semantic topic, we propose an SSN (semantics social network) detection algorithm based on topic comprehensive factor analysis. This algorithm firstly defines the multivariate semantics information as topic, takes multivariate TCF (topic comprehensive factor) as the measurement of topics, and the difference of topic density as polymerization direction, and establishes the initial community structure. Secondly, we establish the cost function with the goal of minimizing the semantics similarity inside the communities and maximizing the semantics similarity between different communities, when some boundary nodes change the community. Thirdly, the SAOP (simulated annealing optimization policy) is established based on the initial community and cost function, which takes the value of cost function as parameter when the boundary nodes change. We could optimize the initial community structure globally and achieve the semantics community detection with multivariate. Finally, the effectiveness of SSN community detection is proved by a serial of simulations.

       

    /

    返回文章
    返回