基于结构聚类挖掘社交网络图
发布时间:2021-05-21 03:06
当今的社会网络,己不再是狭义上社会学研究的内容,转而成为了集尖端的科研价值与巨大的商业潜质于一体的火热研究课题,吸引着愈来愈多各领域的研究人员的关注。随着时代的发展,互联网中的数据也以井喷式的速度急速增加,大数据时代中的网络已经变得异常复杂。随着逐步深入研究复杂网络的物理性质和数学特性,研究者发现许多真实世界的网络除了具备小世界性、无标度性这些特性外,还具有一个共同的特性,那就是社区结构,其由一系列点和边组成,具有社区内部的节点连接十分紧密,社区相互之间的节点连接松散的特征。从社区的角度能更好挖掘网络的功能和价值,且便于分析网络的结构和特性。因而,挖掘出复杂网络中的社区结构具有非常重要的意义。由于缺乏将社交网络转化为数据的有效方法,有权网络和无权网络被当成了两种网络分别研究,大部分对无权网络的算法无法推及至权值网络,基于此,本文主要研究了社交网络转化为数据的方式,使得众多应用于数据的聚类方法可以应用在社交网络上。本文首先简单描述了论文研究的背景、当前的研究现状和本篇论文的组织结构。其次阐述了复杂网络的含义、相关特性、拓扑结构模型、社区的含义,并且描述了几种典型的社区发现算法。以前面的理...
【文章来源】:西安电子科技大学陕西省 211工程院校 教育部直属院校
【文章页数】:76 页
【学位级别】:硕士
【文章目录】:
ABSTRACT
摘要
List of Symbols
List of Abbreviations
Chapter 1 Introduction
1.1 Background
1.2 Research status at home and abroad
1.3 Research content
1.4 Thesis Organization
Chapter 2 Preliminaries
2.1 Graph Theory
2.2 Statistical properties of complex networks
2.2.1 Degree
2.2.2 Neighbors Node
2.2.3 Clustering Coefficient
2.2.4 Shortest Path Length
2.2.5 Density
2.3 The Topology of the Network
2.3.1 Scale-free Network
2.3.2 Small-world Network
2.3.3 Community Structure
2.4 Classic Algorithms
2.4.1 CMP Algorithm
2.4.2 Label Propagation Algorithm
2.4.3 Spectral Clustering
2.4.4 Louvain Algorithm
2.5 Community Evaluation Indicators
2.5.1 Modularity
2.5.2 Normalized Mutual Information
2.5.3 Community in a Strong Sense and in a Weak Sense
2.6 Conclusion
Chapter 3 Community Detection on Pseudo-Adjacency Matrix
3.1 Problem Background and Solution
3.2 K-means Based on Pseudo-adjacency Matrix
3.2.1 The Maximum Degree of Initialization
3.2.2 Basic Idea
3.3 Hierarchical Clustering Based on Pseudo-adjacency Matrix
3.3.1 Similarity Measure Function
3.3.2 Basic Idea
3.4 FCM Algorithm Based on Pseudo-adjacency Matrix
3.4.1 Basic Knowledge
3.4.2 Principle of Algorithm
3.5 Conclusion
Chapter 4 Experimental Results and Analysis
4.1 Data Set
4.1.1 Zachary Karate Club Network
4.1.2 Dolphins Network
4.1.3 Football League Network
4.1.4 Lesmis Network
4.2 K-means Experimental Results and Analysis
4.2.1 Choice of Parameter
4.2.2 Comparative Experiment of Unweighted Social Network
4.2.3 The Experiment of K-means on Weighted Network
4.3 Hierarchical Clustering Experimental Results and Analysis
4.3.1 Choice of Parameter
4.3.2 Comparative Experiment of Unweighted Social Network
4.3.3 Experiment of Hierarchical Clustering on Weighted Network
4.4 FCM Algorithm and Experimental Analysis
4.4.1 Performance Analysis
4.5 Conclusion
Chapter 5 Conclusion and Future Work
5.1 Conclusion
5.2 Future Work
References
Acknowledgements
Biography
【参考文献】:
期刊论文
[1]基于共邻矩阵的复杂网络社区结构划分方法[J]. 郭崇慧,张娜. 系统工程理论与实践. 2010(06)
[2]大型复杂网络中的社区结构发现算法[J]. 胡健,董跃华,杨炳儒. 计算机工程. 2008(19)
本文编号:3198926
【文章来源】:西安电子科技大学陕西省 211工程院校 教育部直属院校
【文章页数】:76 页
【学位级别】:硕士
【文章目录】:
ABSTRACT
摘要
List of Symbols
List of Abbreviations
Chapter 1 Introduction
1.1 Background
1.2 Research status at home and abroad
1.3 Research content
1.4 Thesis Organization
Chapter 2 Preliminaries
2.1 Graph Theory
2.2 Statistical properties of complex networks
2.2.1 Degree
2.2.2 Neighbors Node
2.2.3 Clustering Coefficient
2.2.4 Shortest Path Length
2.2.5 Density
2.3 The Topology of the Network
2.3.1 Scale-free Network
2.3.2 Small-world Network
2.3.3 Community Structure
2.4 Classic Algorithms
2.4.1 CMP Algorithm
2.4.2 Label Propagation Algorithm
2.4.3 Spectral Clustering
2.4.4 Louvain Algorithm
2.5 Community Evaluation Indicators
2.5.1 Modularity
2.5.2 Normalized Mutual Information
2.5.3 Community in a Strong Sense and in a Weak Sense
2.6 Conclusion
Chapter 3 Community Detection on Pseudo-Adjacency Matrix
3.1 Problem Background and Solution
3.2 K-means Based on Pseudo-adjacency Matrix
3.2.1 The Maximum Degree of Initialization
3.2.2 Basic Idea
3.3 Hierarchical Clustering Based on Pseudo-adjacency Matrix
3.3.1 Similarity Measure Function
3.3.2 Basic Idea
3.4 FCM Algorithm Based on Pseudo-adjacency Matrix
3.4.1 Basic Knowledge
3.4.2 Principle of Algorithm
3.5 Conclusion
Chapter 4 Experimental Results and Analysis
4.1 Data Set
4.1.1 Zachary Karate Club Network
4.1.2 Dolphins Network
4.1.3 Football League Network
4.1.4 Lesmis Network
4.2 K-means Experimental Results and Analysis
4.2.1 Choice of Parameter
4.2.2 Comparative Experiment of Unweighted Social Network
4.2.3 The Experiment of K-means on Weighted Network
4.3 Hierarchical Clustering Experimental Results and Analysis
4.3.1 Choice of Parameter
4.3.2 Comparative Experiment of Unweighted Social Network
4.3.3 Experiment of Hierarchical Clustering on Weighted Network
4.4 FCM Algorithm and Experimental Analysis
4.4.1 Performance Analysis
4.5 Conclusion
Chapter 5 Conclusion and Future Work
5.1 Conclusion
5.2 Future Work
References
Acknowledgements
Biography
【参考文献】:
期刊论文
[1]基于共邻矩阵的复杂网络社区结构划分方法[J]. 郭崇慧,张娜. 系统工程理论与实践. 2010(06)
[2]大型复杂网络中的社区结构发现算法[J]. 胡健,董跃华,杨炳儒. 计算机工程. 2008(19)
本文编号:3198926
本文链接:https://www.wllwen.com/kejilunwen/yysx/3198926.html