【作者】 马朝阳；

【导师】 田宏；

【作者基本信息】 大连交通大学 ， 计算机应用技术， 2010， 硕士

【摘要】 社会网络指的是社会行动者及其间的关系集合。也可以说一个社会网络是由多个点(社会行动者)和各点之间的连线(行动者之间的联系)组成的集合。因此社会网络分析不同于单个的语义分析,是注重于关系数据的分析。本文主要围绕应用于网络信息安全的社会网络分析法这一方向展开,研究了基于邮件的社会网络分析,在广泛阅读了国内外文献的基础之上提出了一种基于邮件挖掘社会网络核心层的新方法,本文的主要工作主要有以下两个方面:(1)为了挖掘出完整的社会网络核心层的成员,提出了基于邮件挖掘社会网络核心层的新方法。在用邮件数据构建出了社会网络之后,首先删除节点度小于一定阈值的节点,再运用社团结构挖掘及中心度分析找出部分网络核心成员,最后结合已删除的节点得出完整的网络核心层。实验结果显示,该方法可以找出全部的网络核心成员,且在一定程度上解决了大型网络不容易计算的问题。(2)为了更精确的评价社团结构,本文提出了一种新的评价标准——社团凝聚度,定义社团外部链接数与内部连接数的比值为社团凝聚度,并在此基础上提出了基于局部社团凝聚度增量的社团结构挖掘算法。首先选择初始节点定义为一个社团,然后比较网络中每个节点加入到社团后的凝聚度增量,选择局部社团凝聚度增量增长最快或者减少最慢的节点作为社团成员加入,重复选择合适节点加入社团直到社团凝聚度达到指定阈值,或者发现完全封闭的社团。最后比较挖掘出社团的凝聚度可以确定哪些为社团,哪些则可以作为孤立点。针对本文提出的算法编写程序,应用于计算机生成网络和一个虚拟企业网络,实验结果表明算法是高效的和实用的。更多还原

【Abstract】 Social network refers to the actors and the social relationship between actors. That is to say that a social network is a set composed of a number of points (social actors) and the connection between points (the link between actors). Therefore, social network analysis is different from a single semantic analysis is to focus on the relational data analysis.This article applies to network information security around the social network analysis started in this direction. This paper studies the social network analysis based on e-mail. In this paper, extensive reading at home and abroad based on the literature presents a social network based on excavation e-mail new method for the core layer. The innovation of this paper mainly in the following two aspects:(1)In order to tap the core of complete social network layer members, this paper, mining social networks based on e-mail new method for the core layer. Using e-mail the data to build out a social network, the first delete the node is less than a certain threshold value of the node, then the use of Community Structure Mining and Analysis Center, part of the network to identify the core members of the deleted node last, come to a complete network of core layer. Experimental results show that this method can identify all of the core members of the network, but also to some extent is not easy to solve large-scale network computing problems.(2)In order to more precisely evaluate the degree of polymerization associations proposed in this paper based on the degree of community cohesion of the community structure of incremental mining algorithm. The definition of community External Web Link Counts and internal connections for the community cohesion ratio of the number of degrees, first select the initial node is defined as a society, and then compare the network each node added to the degree of community cohesion after the increment, select local community cohesion degree increments up to the fastest or the slowest node to reduce community members as a repeat selection to find the smallest node join the community until the specified number of members of the community, or find completely closed societies. Comparing the final excavated to determine what degree of community cohesion as a society, and which can be as isolated points.This paper presents an algorithm for programming applied to computer generated networks and a virtual enterprise network, experimental results show that the algorithm is efficient and practical.更多还原