综合显式和隐式关系的社区发现方法

发布时间：2018-06-16 17:21

本文选题：在线社会网络 + 社团发现　；参考：《华中科技大学》2013年硕士论文

【摘要】：在线社会网络是指人们以计算机和网络为媒介进行社交、协作和信息共享所形成的人与人之间的社会网络。近几年，随着Facebook、微博、微信等大规模在线社会网络应用的成功，掀起了在线社会网络的又一轮研究热潮。同社会网络一样，在线社会网络也具有社团结构。探索在线社会网络中的社团结构，可以帮助我们对目前的系统进行评估，并基于它来设计未来的在线社会网络系统，以及帮助系统运营商有效的组织自己的门户。另一方面，社区发现可以帮助产品或服务提供商快速准确的找到对其感兴趣的用户群体，以便于针对性的投放广告信息。本文综合考虑用户的显式和隐式关系，提出一种基于局部模块度的社团发现算法。首先根据在线社会网络应用系统中的用户的行为数据，分析其兴趣爱好，构建用户表示，计算用户之间的相似度，根据一定阀值过滤后，得到用户的隐式关系。然后，根据用户的隐式关系扩展显式关系网络，从而得到一个综合用户显式和隐式关系的网络结构。最后以扩展后得到的网络作为输入，利用基于局部模块度优化的算法，划分在线社会网络的社区结构。本文最后在两个真实数据集上将本文提出的算法与基于兴趣相似度的社团发现算法进行了对比试验，试验结果表明本文提出的综合显式和隐式关系的算法具有更好的社团划分效果。
[Abstract]:The online social network refers to the social network between people which is formed by the social communication, cooperation and information sharing between people with the media of computer and network. In recent years, with the success of Facebook, Weibo, WeChat and other large-scale online social network applications, there has been another wave of research on online social networks. Like social networks, online social networks also have community structures. Exploring the community structure in the online social network can help us to evaluate the current system, design the future online social network system based on it, and help system operators organize their own portal effectively. On the other hand, community discovery can help product or service providers quickly and accurately find the user groups interested in them, so as to facilitate targeted advertising information. In this paper, a community discovery algorithm based on local modularity is proposed considering the explicit and implicit relationships of users. Firstly, according to the behavior data of the users in the online social network application system, we analyze their interests and hobbies, construct the user representation, calculate the similarity between users, filter according to a certain threshold, and get the implicit relationship of the users. Then, the explicit relational network is extended according to the implicit relation of the user, and a network structure combining the explicit and implicit relationship between the user and the user is obtained. Finally, the extended network is used as the input, and the community structure of the online social network is divided by using the algorithm based on local modularity optimization. In the end of this paper, we compare the proposed algorithm with the community discovery algorithm based on interest similarity on two real data sets. The experimental results show that the proposed algorithm of combining explicit and implicit relations has a better effect on community division.
【学位授予单位】：华中科技大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：O157.5;TP393.09

【参考文献】