摘要
如何准确、有效地发现虚拟社交网络中的社区或群体是复杂社交网络中的一个热点问题。本研究认为在线社交网络中用户之间显性的对话或彼此评论形成了一种网络结构,既包含社交网络底层的拓扑结构信息,又包含网络实体进行交互的确切时间,具有时效性。为了揭示虚拟社交网络中隐藏的动态现象,给出了一种同时考虑主题和时间的在线社交网络发现算法——多时间密集子图发现算法。首先,将在线社交网络中的对话或评论建模为一个交互网络,再利用拓扑结构将网络划分为属于不同主题(如热门新闻或话题)的社区,然后依据时间维度对每个主题下的社区挖掘稠密子图,最后在真实数据集上对所提算法进行了评估。与比较算法相比,本算法发现的团体内部间的交互在时间上更加密集,具有更高的时效性。
The accurate and effective discovery of communities or groups in virtual social networks is a hot issue in complex social networks.The explicit conversations or comments between users in an online social network form a network structure that contains information about the underlying topology of the social network and the exact time when network entities interact,and is thus time-dependent.In order to reveal the hidden dynamic phenomena in virtual social networks,this paper presents an online social network discovery algorithm(multiple temporal dense subgraph discovery algorithm,MTDSDA)considering both themes and time.Firstly,conversations or comments on online social networks were modeled intoan interactive network.Secondly,the topology structure was adopted to divide the network into communities with different themes(such as popular news or topic).Thirdly,dense subgraphs for the communities under each theme were mined by considering the time dimension.Finally,the proposed algorithm was evaluated on the real-worlddatasets.Compared with the comparison algorithm,the interaction between the groups discovered by the algorithm is more intensive in time and has higher chronergy.
作者
李金鹏
曹宁
张琪
张文鹏
纪淑娟
LI Jinpeng;CAO Ning;ZHANG Qi;ZHANG Wenpeng;JI Shujuan(Shandong Provincial Key Laboratory of Wisdom Mine Information Technology, Shandong University of Science and Technology, Qingdao, Shandong 266590, China)
出处
《山东科技大学学报（自然科学版）》
CAS
北大核心
2021年第4期94-102,共9页
Journal of Shandong University of Science and Technology(Natural Science)
基金
国家自然科学基金项目(71772107)
青岛社会科学规划研究项目(QDSKL1801138)
山东省自然科学基金项目(ZR2018BF013)
山东省研究生质量提升计划项目(2016)
山东科技大学领军人才计划项目(2014)
山海英才计划项目(2019)
泰山学者攀登计划项目(2014)。
关键词
在线社交网络
图挖掘
主题划分
动态密集子图
团体发现
online social network
graph mining
division ofthemes
dynamic dense subgraph
groups discovery
