一种面向语义重叠社区发现的 Block 场取样算法

语义社会网络(Semantic social network, SSN)是一种包含信息节点及社会关系构成的新型复杂网络. 传统语义社会网络分析算法在进行社区挖掘时, 需要预先设定社区个数且无法发现重叠社区. 针对这一问题, 提出一种面向语义重叠社区发现的block场采样算法, 该算法首先以LDA (Latent dirichlet allocation)模型为语义分析模型, 建立了以取样节点为核心节点的block 场BAT (Block-author-topic)模型; 其次, 根据节点的语义分析结果, 建立可度量block区域的语义凝聚力方法, 实现了语义信息的可度量化; 最后, 以节点的语义凝聚力为输入, 改进了重叠社区发现的标签传播算法(Label propagation algorithm, LPA)及可评价语义社区的SQ度量模型, 并通过实验分析, 验证了本文算法及SQ 度量模型的有效性及可行性.

An Overlapping Community Structure Detecting Algorithm in Semantic Social Network Based on Block Field

The semantic social network (SSN) is a new kind of complex networks consisting of the node content and topological relationship. The traditional community detection algorithms need to preset the number of the communities and could not detect the overlapping communities. To solve this problem, an overlapping community structure detecting algorithm in semantic social network based on the block field is proposed. Firstly, it takes the latent dirichlet allocation (LDA) model as the semantic analyzing model, establishing the block-author-topic (BAT) model with the sampling node as the core node. Secondly, it suggests the measurement of the semantic cohesion of the block field, depending on the analysis of SSN, to achieve the evaluation of semantic information. Finally, it improves the label propagation algorithm (LPA) which could detect the overlapping communities, with the semantic cohesion as input, and designs the SQ measurement modularity for semantic measuring. The efficiency and feasibility of the algorithm and the semantic modularity are verified via experimental analysis.