基于话题概率模型的语义社区发现方法研究

语义社会网络(Semantic social network, SSN)是一种由信息节点及社会关系构成的复杂网络, 也是语义信息时代社会网络技术研究的热点, 相较于传统社会网络更具实用价值. 其研究内容包含了社会网络的语义分析及社会关系分析, 因此, 语义社会网络的社区挖掘建模具有一定的复杂性. 在语义社会网络的社区挖掘研究方面, 本文分析了当前基于话题概率模型的语义社区发现方法, 并在综述其内容的同时总结了各方法的优缺点, 为后续研究提供了理论基础. 在语义社会网络社区挖掘结果的评判方面, 本文归纳了相关的评价模型, 并通过实验分析对比了各模型对拓扑相关性和语义相关性的倾向性.

Semantic Community Detection Research Based on Topic Probability Models

Semantic social network (SSN) is a complex network consisting of textual node information and social relationships, and has more valuable applications than the traditional social network which focuses on social network analysis (SNA) with semantic information. As it contains the semantic analysis and social relationship analysis, there is some complexity for the modeling in mining the SNA community. In the SNA community mining aspect, the advantages and disadvantages of each method based on topic probability models are summarized to provide a theoretical basis for the further research. In the evaluation aspect of SNA community mining, relevant evaluation models are summarized. The tendency of each evaluation model toward topological and semantic relevance is compared by experimental analysis.