基于多目标优化的符号网络社区检测算法

符号网络可以描述实体之间的多种关系,对符号网络中的社团检测可以挖掘出其中的有效信息.同时考虑连接密度和连接符号,将社团发现问题建模为一个多目标优化问题,基于MOEA/D框架,提出一种改进的符号网络社团发现算法,设计了基于字符串的编码方式、预分区策略、交叉合并策略、变异方式等.实验结果表明,本算法可以有效检测出社团结构.

SIGNED NETWORK COMMUNITY DISCOVERY ALGORITHM BASED ON MULTI-OBJECTIVE OPTIMIZATION

The signed network can describe a variety of relationships between entities, and the community detection in the signed network can dig out the effective information. At the same time, considering connection density and connection symbols, the community discovery problem is modeled as a multi-objective optimization problem. Based on the MOEA/D framework, an improved signed network community discovery algorithm is proposed, and a string-based encoding method and pre-partitioning strategy are designed. Also cross-merger strategy, mutation method are designed. The experimental results show that the algorithm can effectively detect the community structure.