基于灰色方法的社交网络群体识别问题研究

项目名称： 基于灰色方法的社交网络群体识别问题研究

项目编号： No.61300104

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 自动化技术、计算机技术

项目作者： 郭昆

作者单位： 福州大学

项目金额： 23万元

中文摘要： 社交网络在世界范围内的飞速发展正引起越来越多学者的关注，由于其具有超大规模性、重叠性、层次性、多重性、动态性等不同于一般复杂网络的特性，使社交网络中的群体识别面临许多挑战。本研究拟在灰色系统理论的指导下，以灰色方法为主要工具，通过与包括概率统计方法、模糊方法等在内的多种不确定方法相结合，建立具有重叠、层次等复杂结构的社交网络、动态社交网络及具有复杂结构的动态社交网络的灰色模型。同时，引入进化计算方法及并行计算框架，设计具有强鲁棒性、高扩展性及可并行化的高效算法，通过理论研究与仿真实验相结合的方式对算法的性能展开评价与分析。研究的成果将在理论上从新的角度为复杂网络聚类方法提供有益补充，同时拓展灰色系统理论及灰色方法在新的学科领域的应用，在实践上对促进基于社交网络的新型社会、经济活动的发展、识别与预防基于社交网络的犯罪行为、维护国家安全等都具有重要的现实意义。

中文关键词： 社交网络；社区发现；灰色方法；复杂网络；聚类

英文摘要： The rapid development of the social networks has been attracting the attention of more and more scholars. However, the social networks possess many spectific characteristics like extremely large scale, overlapping, multi-level structures, multiple relations, dynamicity, etc, some of which are lacked in the general complex networks. The specifities present great challenges to the community detection in the social networks. In this research project, guided by the grey system theory and based on the grey methods, we try to build grey models for the social networks with overlapping, multi-level and other complex structures and dynamic social networks with multiple uncertainty methods including statistics methods, fuzzy methods, etc. Then, efficient algorithms with strong robustness and high scalibility are developed through introducing the evolutionary computation methods and parallel computing frameworks. The performance of the algorithms are evaluated by theoretical analysis and simulated experiments. In theory, the results of the research will provide beneficial complement to the existing complex network clustering methods from a new perspective and boost the application of grey system theory and grey methods in new disciplines and realms. In pratice, the research results are important to the promotion of the dev

英文关键词： social network；community discovery；grey methods；complex network；clustering