基于随机块模型的大规模社会网络中观关键结构研究

发布时间：2018-01-09 02:27

  本文关键词：基于随机块模型的大规模社会网络中观关键结构研究　出处：《太原理工大学》2015年硕士论文　论文类型：学位论文

  更多相关文章： 大规模社会网络 关键结构 随机块模型 社区结构 结构洞节点

【摘要】：随着互联网的不断发展，智能手机、平板电脑等智能终端在人类生活中的普及以及移动网络带宽的不断提高，使得微信、微博等社交媒体已经渐渐成为了人类生活中不可或缺的部分。人们在真实世界中的互动与联系，也不断的向互联网上拓展，使得我们的社会也被社会化了。作为一种社会学理论、信息处理及电子商务等多个学科的交叉热点，大规模社会网络应运而生。大规模社会网络中，每天产生的海量信息数据具有很高的价值。对这些数据的发掘和分析成为研究大规模社会网络的重要途径。通过对大规模社会网络数据的分析和有效信息的发掘，能够更好的理解大规模社会网络，为进一步的舆情监控、电子商务个性化推荐等实际应用的研究提供了理论基础。 目前，社会网络的研究已经逐渐从对中小规模的在线社会网络结构研究，进入到了大规模社会网络结构的研究中。同时，研究的目标也从发现指定单一结构方面向多结构发现方向转变。大规模社会网络结构多样化研究已经成为了当前社会网络结构研究的热点和重点。然而，现有的研究，往往从特定的结构研究出发，没有考虑到多种结构相互配合，能够更好的反应出大规模社会网络结构的特点。基于以上的考虑，本文针对大规模社会网络的结构问题进行了研究，提出了新的概念及相关定义与实现算法。本文的贡献主要有以下几方面： 首先，在大规模社会网络结构研究的传统定义上，，充分考虑了多种结构在大规模社会网络结构分析的重要性，提出了关键结构概念。同时分别给出了关键结构概念的一般化定义。并依照关键结构的定义，说明了关键结构的发现过程，提出了关键结构的评价标准。 其次，给出了结构洞节点的描述性定义，分别介绍了用于发现关键结构的随机块模型SBM(Stochastic Block Model)及结构洞节点发现方法SHSD(Structural Hole Spanners Detection)。 最后，使用名为BC的博客数据集和名为MB的微博数据集为实验数据集，进行相关实验。实验结果表明，随机块模型能够发现社会网络中的社区结构，且在分类结果上使用结构洞发现方法SHSD能够发现网络中的结构洞节点。通过与原始输入网络数据图进行比较，实验发现的关键结构能够较为全面的描述整个社会网络的结构特征。
[Abstract]:With the continuous development of the Internet, smart phones, tablets and other intelligent terminals in human life, as well as mobile network bandwidth continues to improve, making WeChat. Social media such as Weibo have become an integral part of human life. People's interactions and connections in the real world are expanding to the Internet. As a kind of sociological theory, information processing and electronic commerce and so on, the large-scale social network emerges as the times require. In the large-scale social network, the large-scale social network emerges as the times require. The massive information data produced every day is of great value. The discovery and analysis of these data has become an important way to study large-scale social network. Through the analysis of large-scale social network data and effective information mining. . It can better understand the large-scale social network, which provides a theoretical basis for the further research of public opinion monitoring, e-commerce personalized recommendation and other practical applications. At present, the research of social network has gradually moved from the research of small and medium scale online social network structure to the research of large-scale social network structure. At the same time. The goal of the study is also changed from finding a single structure to a multi-structure discovery. The research of large-scale social network structure diversification has become the focus and focus of the current social network research. Existing studies, often from the specific structure of the study, do not take into account a variety of structures to cooperate with each other, can better reflect the characteristics of large-scale social network structure. Based on the above considerations. In this paper, the structure of large-scale social networks is studied, and a new concept, related definitions and implementation algorithms are proposed. The contributions of this paper are as follows: Firstly, in the traditional definition of large-scale social network structure research, the importance of various structures in large-scale social network structure analysis is fully considered. In this paper, the concept of key structure is put forward, and the general definition of the concept of key structure is given. According to the definition of key structure, the discovery process of key structure is explained, and the evaluation criteria of key structure are put forward. Secondly, the descriptive definition of the structure hole node is given. The random block model (SBM(Stochastic Block Model) used to discover key structures and the method of structural hole node discovery (SHSD) are introduced respectively. Structural Hole Spanners detection. Finally, using the blog data set named BC and Weibo dataset named MB as the experimental data set, the experiment results show that the random block model can find the community structure in the social network. The structural hole nodes in the network can be found by using the structure hole discovery method (SHSD) in the classification results, and compared with the original input network data graph. The key structure found in the experiment can describe the structural characteristics of the whole social network comprehensively.
【学位授予单位】：太原理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP393.092;TP311.13

【参考文献】

中国期刊全文数据库 前6条

1 陈婷婷;;结构洞:关系的制胜点[J];三峡大学学报(人文社会科学版);2007年S1期

2 窦炳琳;李澍淞;张世永;;基于结构的社会网络分析[J];计算机学报;2012年04期

3 王莉;程苏琦;沈华伟;程学旗;;在线社会网络共演化的结构推断与预测[J];计算机研究与发展;2013年12期

4 徐恪;张赛;陈昊;李海涛;;在线社会网络的测量与分析[J];计算机学报;2014年01期

5 刘军;;法村社会支持网络的整体结构研究块模型及其应用[J];社会;2006年03期

6 柴变芳;于剑;贾彩燕;王静红;;一种基于随机块模型的快速广义社区发现算法[J];软件学报;2013年11期

本文编号：1399659