面向微博用户的社交网络社区发现研究
发布时间：2018-11-23 10:41
【摘要】：随着计算机技术快速发展,社交网络应运而生改变了人们面对面的交流方式,由传统的线下沟通变革为新时代的“线上”沟通及“掌心”交流。如今处于大数据时代,社交网络中海量数据对于社会科学研究显得更加重要,而发现社交网络的社区结构已然成为学者们研究的热点领域。社区发现技术对研究复杂网络拓扑结构有很大的帮助,同时也蕴含很大的社会价值。目前在社区发现技术在复杂网络领域已经取得了不错的成果,但针对社交网络的社区发现技术还不太成熟,因为社交网络规模较大且内容繁杂,大多算法都存在一定的缺陷,如算法复杂度过高,结果不够准确或局部最优等。鉴于此,本文针对社交网络中的微博网络进行研究,从微博用户出发,通过用户关系和用户内容的融合,发现潜在的用户社区,并且通过实验证实了结果的合理性。本文主要做了以下几方面的研究和创新:(1)针对微博网络中用户关系结构的特点,考虑到网络中同时存在单向关注和双向关注两类关系,提出了一种计算用户关系相似度的方法,该方法兼顾这两类关注关系对节点的影响,同时将有向网络转换为加权无向网络进行计算,提高了运行效率。另外针对加权无向网络,利用用户相似度作为权重提出了一种改进的CNM社区发现算法。根据朋友的朋友更容易成为朋友的思想,可以延伸为朋友的朋友和自己同相似,所以用节点相似度替代模块度进行社区合并,更加合理的发现用户社区。这是针对网络中用户的关系特点进行社区发现。(2)微博网络中用户内容可以反映用户当前的兴趣,针对这一思想,提出了用户关系和用户内容融合的社区发现算法。根据主题模型的思想融入用户标签来发现用户的兴趣主题,通过相对熵计算用户兴趣主题的相似度,同时加入用户关系相似度并通过实验调节两类相似度融合的比重,充分体现用户的兴趣特性。(3)在融合用户关系和内容两种相似度的基础上,提出了JSCNM算法,利用改进模块度增量函数将融合后的中心度加入到优化函数中,充分考虑微博网络中关系和内容对节点影响力作用,经过不断寻找最优目标达到划分社区目的。利用微博网络真实数据集进行实验,结果证明划分社区更加合理。
[Abstract]:With the rapid development of computer technology, social networks have changed the way people communicate face to face, from the traditional offline communication to the new era of "online" communication and "palm" communication. In the era of big data, the mass data in social network is more important for social science research, and the discovery of social network community structure has become a hot research area of scholars. Community discovery technology is of great help to the study of complex network topology, and also has great social value. At present, the community discovery technology has made good achievements in the field of complex network, but the community discovery technology for social network is not very mature, because the social network is large and complex, most algorithms have some defects. If the complexity of the algorithm is too high, the results are not accurate or local optimal. In view of this, this paper focuses on Weibo network in social network. From the perspective of Weibo users, through the fusion of user relationship and user content, the potential user community is discovered, and the rationality of the result is verified by experiments. This paper mainly makes the following aspects of research and innovation: (1) considering the characteristics of user relationship structure in Weibo network, considering that there are two kinds of relationships in the network: unidirectional concern and two-way concern, This paper presents a method to calculate the similarity of user relationships. This method takes into account the influence of these two kinds of relationships on nodes, and transforms the directed network into a weighted undirected network for computation, which improves the running efficiency. In addition, an improved CNM community discovery algorithm based on user similarity is proposed for weighted undirected networks. According to the idea that a friend of a friend is more likely to be a friend, it can be extended to a friend who is similar to himself, so the node similarity degree is used instead of the module degree to merge the community and to find the user community more reasonably. This is based on the characteristics of user relationships in the network. (2) user content in Weibo network can reflect the current interests of users. In view of this idea, a community discovery algorithm of user relationship and user content fusion is proposed. According to the idea of topic model, the user's topic of interest is found by integrating the idea of topic model into user's label, and the similarity of user's topic of interest is calculated by relative entropy. At the same time, the similarity of user relationship is added and the proportion of the fusion of two kinds of similarity is adjusted by experiment. (3) on the basis of merging user relationship and content similarity, JSCNM algorithm is proposed, and the improved modularity increment function is used to add the merged centrality to the optimization function. Considering the influence of the relationship and content on the nodes in Weibo's network, the goal of dividing the community is achieved by searching for the optimal goal. Using Weibo network real data set to experiment, the results show that the division of community is more reasonable.
【学位授予单位】：江西理工大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP393.09
[Abstract]:With the rapid development of computer technology, social networks have changed the way people communicate face to face, from the traditional offline communication to the new era of "online" communication and "palm" communication. In the era of big data, the mass data in social network is more important for social science research, and the discovery of social network community structure has become a hot research area of scholars. Community discovery technology is of great help to the study of complex network topology, and also has great social value. At present, the community discovery technology has made good achievements in the field of complex network, but the community discovery technology for social network is not very mature, because the social network is large and complex, most algorithms have some defects. If the complexity of the algorithm is too high, the results are not accurate or local optimal. In view of this, this paper focuses on Weibo network in social network. From the perspective of Weibo users, through the fusion of user relationship and user content, the potential user community is discovered, and the rationality of the result is verified by experiments. This paper mainly makes the following aspects of research and innovation: (1) considering the characteristics of user relationship structure in Weibo network, considering that there are two kinds of relationships in the network: unidirectional concern and two-way concern, This paper presents a method to calculate the similarity of user relationships. This method takes into account the influence of these two kinds of relationships on nodes, and transforms the directed network into a weighted undirected network for computation, which improves the running efficiency. In addition, an improved CNM community discovery algorithm based on user similarity is proposed for weighted undirected networks. According to the idea that a friend of a friend is more likely to be a friend, it can be extended to a friend who is similar to himself, so the node similarity degree is used instead of the module degree to merge the community and to find the user community more reasonably. This is based on the characteristics of user relationships in the network. (2) user content in Weibo network can reflect the current interests of users. In view of this idea, a community discovery algorithm of user relationship and user content fusion is proposed. According to the idea of topic model, the user's topic of interest is found by integrating the idea of topic model into user's label, and the similarity of user's topic of interest is calculated by relative entropy. At the same time, the similarity of user relationship is added and the proportion of the fusion of two kinds of similarity is adjusted by experiment. (3) on the basis of merging user relationship and content similarity, JSCNM algorithm is proposed, and the improved modularity increment function is used to add the merged centrality to the optimization function. Considering the influence of the relationship and content on the nodes in Weibo's network, the goal of dividing the community is achieved by searching for the optimal goal. Using Weibo network real data set to experiment, the results show that the division of community is more reasonable.
【学位授予单位】：江西理工大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP393.09
【相似文献】
相关期刊论文 前10条
1 ;基于位置的手机社交网络“贝多”正式发布[J];中国新通信;2008年06期
2 曹增辉;;社交网络更偏向于用户工具[J];信息网络;2009年11期
3 ;美国:印刷企业青睐社交网络营销新方式[J];中国包装工业;2010年Z1期
4 李智惠;柳承烨;;韩国移动社交网络服务的类型分析与促进方案[J];现代传播(中国传媒大学学报);2010年08期
5 贾富;;改变一切的社交网络[J];互联网天地;2011年04期
6 谭拯;;社交网络:连接与发现[J];广东通信技术;2011年07期
7 陈一舟;;社交网络的发展趋势[J];传媒;2011年12期
8 殷乐;;全球社交网络新态势及文化影响[J];新闻与写作;2012年01期
9 许丽;;社交网络:孤独年代的集体狂欢[J];上海信息化;2012年09期
10 李玲丽;吴新年;;科研社交网络的发展现状及趋势分析[J];图书馆学研究;2013年01期
相关会议论文 前10条
1 赵云龙;李艳兵;;社交网络用户的人格预测与关系强度研究[A];第七届（2012）中国管理学年会商务智能分会场论文集（选编）[C];2012年
2 宫广宇;李开军;;对社交网络中信息传播的分析和思考——以人人网为例[A];首届华中地区新闻与传播学科研究生学术论坛获奖论文[C];2010年
3 杨子鹏;乔丽娟;王梦思;杨雪迎;孟子冰;张禹;;社交网络与大学生焦虑缓解[A];心理学与创新能力提升——第十六届全国心理学学术会议论文集[C];2013年
4 毕雪梅;;体育虚拟社区中的体育社交网络解析[A];第九届全国体育科学大会论文摘要汇编(4)[C];2011年
5 杜p，
本文编号：2351295
本文链接：https://www.wllwen.com/guanlilunwen/ydhl/2351295.html
