本文已被:浏览 4725次 下载 7643次
中文摘要: 当前我国 互联网信息安全形势严峻, 在社会舆情预警、隐私泄露保护等方面具有重
大需求。 在线社交网络由于其承载着海量的信息,是信息安全研究的重要对象。 本文从作为
理论基础的社交网络结构属性和模型展开综述, 其中包含作者所在研究团队的成果及观点。
社交网络的结构特性的研究重点在于揭示社交网络的基本属性、探索作为社交网络基本要素
的人与人之间的关系的本质特征;社交网络模型研究重点在于通过模拟真实社交关系网中人
与人之间的交往行为来构造具备相应属性特征的演化模型, 并以此为基础, 研究特定社交行
为对网络结构的影响, 或者通过所建立的模型逆向分析社交网络的本质特征、推断哪些社交
行为决定着相应的网络特性。 近年来, 不同的策略和技术分别在网络的结构特征、网络中的
信息传播和检索、网络中用户的行为分析、社团结构挖掘等领域取得了重大研究进展。 文章
对网络结构研究中取得的成果进行了回顾,分别介绍了近期研究中所发现的社交网络在不同
层面表现出的结构特征,并给出了社交网络模型研究中几类常见的结构模型和建模方法。
中文关键词: 社交网络,网络结构,结构特征,结构建模
Abstract:With the rapid growth of social network services, social network has been an important scientific research area. Researches on network structure, information diffusion and retrieval, user behavior analysis, community detection on social network have made great progress. In this article, domestic and international researches progress on social network structure is reviewed. We focus on two main issues of this research direction, namely, feature analysis and structural modelling of social networks. In the end, we summarize the existing problems and give some possible solutions for enlightening the researchers on this area and for providing
guidance for government’ s decision making.
The study of networks in form of graph theory is one of the fundamental ontologies to understand the nature behind networks. The term structure is named by using the same term in the domain of graph theory, which
implies vertices, edges, and the connection between them. In order to comprehend the mechanism of social network formation and the laws behind the social behavior, we should always study the structure of social network
in prior. By studying network structure, we can also gain deeper insights on other specific researching fields on
social network, such as information retrieval, influence maximization, user behavior analysis, recommendation,
etc.
The topic of research on social network structure consists of two main parts: structural feature analysis and
network structure modelling. Analyzing the features of social network is a necessary preparation for modelling
social network structure. In this article, we give a comprehensive review of structural features and properties of
social network first. Then, two types of models about social network structure are introduced. In the part of feature analysis of social network, we enumerate major features of social network, such as“small world” phenomenon, scale free property, centrality, network resilience, assortative mixing, and community structure. Then, we
introduce one excellent work about the evolution of social network structure. In this work, it is confirmed that
due to the universal cooperation, the structural characteristics in modern social network and online social network have long been existed in the ancient times. In the part of modelling social network structure, we firstly introduce the type of basic generative model. These models originated from the ER random networks model and
developed into a large family of network models featured by“small-world” phenomenon and scale free properties. By simulating human behavior in online social networks, several models are proposed and they are capable of recovering structural features of the observed network data in different aspects. Secondly, the type of stochastic generative models are introduced, these kinds of models can be used to postulate complex latent structures responsible for a set of network observations, and this makes it possible to recover this structure by statistical inference. It is also emphasized that in order to fit the data well, one need to guard against that the proposed model is realistic enough.
Future research directions are discussed in the last section. One important direction is to prove the coritivity conjecture.
The coritivity conjecture says that community structures are formed by minimizing the coritivity of its subgraph and the feature graph of social network, in which by replacing all communities with equal number of nodes, the derived network also follows the principle of coritivity minimization. The verification and proof of this conjecture will solve some vital problems in the
domain of social network, such as understanding mechanisms of structural formation. Other important future directions include
introducing genus from algebraic structure, defining community structures, quantification of individual rationality. We hope
this paper will provide inspirations for researchers and an insight for the government’ s network security strategy in the view of
network science.
文章编号: 中图分类号: 文献标志码:
基金项目:
作者 | 单位 |
许进 |
Author Name | Affiliation |
Xu Jin |
引用文本:
许进.[2015年第2期] 社交网络结构特性分析及建模研究进展[J].中国科学院院刊,2015,30(2):.
Xu Jin.Social network structure feature analysis and its modelling[J].Bulletin of Chinese Academy of Sciences,2015,30(2):.
许进.[2015年第2期] 社交网络结构特性分析及建模研究进展[J].中国科学院院刊,2015,30(2):.
Xu Jin.Social network structure feature analysis and its modelling[J].Bulletin of Chinese Academy of Sciences,2015,30(2):.