基于社群智能的大规模移动社交网络数据挖掘技术研究  

Research on Mining Large-scale Mobile Social Network Data Based on Social and Community Intelligence

作　　者：魏军林 韩楠[2] 乔少杰 袁犁 丁超 WEI Jun-Lin;HAN Nan;QIAO Shao-Jie;YUAN Li;DING Chao(Sichuan JKC Geographical Information Technologies Co.,Ltd.,Chengdu,610043;School of Cybersecurity,Chengdu University of Information Technology,Chengdu,610225;School of Management,Chengdu University of Information Technology,Chengdu,610103;Panzhihua City Geographic Information Center,Panzhihua,617000)

机构地区：[1]四川省金科成地理信息技术有限公司,四川成都610043 [2]成都信息工程大学管理学院,四川成都610103 [3]成都信息工程大学网络空间安全学院,四川成都610225 [4]攀枝花市地理信息中心,四川攀枝花617000

出　　处：《数码设计》2018年第6期1-3,共3页Peak Data Science

基　　金：四川省科技计划资助(2018JY0448);国家自然科学基金(61772091,61802035);四川高校科研创新团队建设计划(18TD0027);成都信息工程大学科研基金(KYTZ201715,KYTZ201750);成都信息工程大学中青年学术带头人科研基金(J201701);广东省普通高校国家级重大培育项目(2014GKXM054);四川省教育厅资助科研项目(14ZB0458);成都市软科学研究项目(2017-RK00-00053-ZF)资助.

摘　　要：将大规模移动社交网络数据作为研究对象,以社群智能、数据挖掘、GIS为主要手段,对网络宏观静态拓扑特性和微观交互规律进行研究。从模型通用性入手,对个体交互模式建模,揭示群体行为特征,为社会计算、现实世界挖掘、城市计算研究奠定基础。研究难点和关键问题包括:(1)利用复杂网络拓扑特性理论研究网络直径缩减属性和稠化定律;(2)设计基于时空相似性的个体和群体交互模式发现算法;(3)设计新型层次聚类算法解决社区发现问题。切实提高利用社群智能理论对不同类型和规模移动社交网络进行数据挖掘和知识发现的准确率和效率。In order to analyze large-scale mobile social network data, this study employs the social and community intelligence, data mining and GIS techniques to explore the statically macroscopic topological properties and the dynamically microscopic interaction rules. In order to design a general knowledge discovery schema, this paper models the individual interactions, discloses the characteristic of social behavior, which can help set up the methodology on social computing, reality mining, and urban computing. The difficulties and essential problems in this paper include:(1) using the topological property analysis theories in complex networks to evaluate the shrinking diameter property and densification law;(2) proposing novel individual and social interaction discovery algorithms based on spatio-temporal similarity of individuals;and (3) proposing new hierarchical clustering approaches to cope with the problem of community discovery. This study can greatly help improve the effectiveness and efficiency of data mining and knowledge discovery in different kinds of mobile social networks with various sizes based on the social and community intelligence theories.

关 键 词：移动社交网络 社区智能 拓扑结构分析 社区发现 层次聚类 GIS 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]