基于交互度的大规模社会网络社区发现研究
发布时间：2019-05-10 21:43
【摘要】：近几年随着互联网“爆炸性”的增长,人与人就交流得更加密切,社会网络也随之飞速发展,而社会网络又是由个人和组织以及他们之间的联系(交互)构成的集合,而现在研究的热点问题也逐渐变成了大规模的社会网络发现问题。在现实生活中,网络是与我们息息相关的,社会的政治经济活动,朋友之间的日常交际,科学家的合作等这些现象可以从社会网络的角度研究。通过研究这样的网络结构,我们可以用于分析信息的传播规律,把有着共同兴趣爱好的人们连接在一起,为人们提供一个更完善的交流平台和工具,或阻止不良信息的传播。本文就是基于人与人之间的交互行为,提出交互度用来度量网络成员之间的交互,并基于交互度提出了一个大规模社会网络的社区发现算法。本文的主要工作概括如下： (1)提出分层次大规模网络社区发现的方法,大规模网络的核心问题是网络规模过大,针对这个问题,本文利用分层聚类的思想提出了大规模社会网络的社区发现算法。首先,将大规模网络先进行预处理,划分为局部小的网络,并在小规模网络中进行社区发现。随后,将小规模网络中发现的社区看作一个点,在后处理阶段重构网络,基于重构网络进行全局社区发现。 (2)基于交互度的计算,完成了交互度量从局部向全局的传递。大规模网络划分以后,通过交互度的计算,成员之间和社区之间的交互关系度量能够从局部向全局传递,保持度量的一致性,从而保证社区发现结果的准确性。 (3)算法被应用于真实社会网络和人工模拟的网络上,分别对本文算法的准确度和效率两个方面进行了测试,说明了算法有效性和效率。 综上所述,本文的工作是面对社会网络分析的背景下,提出基于交互度的大规模社会网络的社区发现算法,并且能够得到一个较为精确的社区划分。
[Abstract]:In recent years, with the "explosive" growth of the Internet, people have been communicating more closely with each other, and social networks have also developed rapidly, and social networks are a collection of individuals and organizations and their connections (interactions). And now the hot issue of research has gradually become a large-scale social network discovery problem. In real life, the network is closely related to us. The social political and economic activities, the daily communication between friends, the cooperation of scientists and so on can be studied from the perspective of social network. By studying such a network structure, we can be used to analyze the law of information dissemination, connect people with common interests, provide people with a more perfect communication platform and tools, or prevent the dissemination of bad information. In this paper, based on the interaction behavior between people, the interaction degree is proposed to measure the interaction between network members, and a community discovery algorithm for large-scale social networks is proposed based on the interaction degree. The main work of this paper is summarized as follows: (1) the method of hierarchical large-scale network community discovery is proposed. The core problem of large-scale network is that the network scale is too large. In this paper, a community discovery algorithm for large-scale social networks is proposed by using the idea of hierarchical clustering. First of all, the large-scale network is preprocessed and divided into locally small networks, and community discovery is carried out in small-scale networks. Then, the community found in the small-scale network is regarded as a point, and the network is reconstructed in the post-processing stage, and the global community discovery is carried out based on the reconstructed network. (2) based on the calculation of interaction degree, the transfer of interaction measurement from local to global is completed. After large-scale network partition, through the calculation of interaction degree, the measurement of interaction between members and communities can be transferred from local to global, and the consistency of measurement can be maintained, so as to ensure the accuracy of community discovery results. (3) the algorithm is applied to real social network and artificial simulated network, and the accuracy and efficiency of the algorithm are tested respectively, and the effectiveness and efficiency of the algorithm are illustrated. To sum up, the work of this paper is to propose a large-scale social network community discovery algorithm based on interaction under the background of social network analysis, and can get a more accurate community division.
【学位授予单位】：云南大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：O157.5
本文编号：2474022
[Abstract]:In recent years, with the "explosive" growth of the Internet, people have been communicating more closely with each other, and social networks have also developed rapidly, and social networks are a collection of individuals and organizations and their connections (interactions). And now the hot issue of research has gradually become a large-scale social network discovery problem. In real life, the network is closely related to us. The social political and economic activities, the daily communication between friends, the cooperation of scientists and so on can be studied from the perspective of social network. By studying such a network structure, we can be used to analyze the law of information dissemination, connect people with common interests, provide people with a more perfect communication platform and tools, or prevent the dissemination of bad information. In this paper, based on the interaction behavior between people, the interaction degree is proposed to measure the interaction between network members, and a community discovery algorithm for large-scale social networks is proposed based on the interaction degree. The main work of this paper is summarized as follows: (1) the method of hierarchical large-scale network community discovery is proposed. The core problem of large-scale network is that the network scale is too large. In this paper, a community discovery algorithm for large-scale social networks is proposed by using the idea of hierarchical clustering. First of all, the large-scale network is preprocessed and divided into locally small networks, and community discovery is carried out in small-scale networks. Then, the community found in the small-scale network is regarded as a point, and the network is reconstructed in the post-processing stage, and the global community discovery is carried out based on the reconstructed network. (2) based on the calculation of interaction degree, the transfer of interaction measurement from local to global is completed. After large-scale network partition, through the calculation of interaction degree, the measurement of interaction between members and communities can be transferred from local to global, and the consistency of measurement can be maintained, so as to ensure the accuracy of community discovery results. (3) the algorithm is applied to real social network and artificial simulated network, and the accuracy and efficiency of the algorithm are tested respectively, and the effectiveness and efficiency of the algorithm are illustrated. To sum up, the work of this paper is to propose a large-scale social network community discovery algorithm based on interaction under the background of social network analysis, and can get a more accurate community division.
【学位授予单位】：云南大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：O157.5
【参考文献】
相关期刊论文 前2条
1 王刚;钟国祥;;基于信息熵的社区发现算法研究[J];计算机科学;2011年02期
2 王林,戴冠中;复杂网络中的社区发现——理论与应用[J];科技导报;2005年08期
，本文编号：2474022
本文链接：https://www.wllwen.com/jingjilunwen/zhengzhijingjixuelunwen/2474022.html
