Development and Application of a Directional Link Analysis Technique

• Main Authors: Sheng-Wei Huang, 黃勝偉
• Other Authors: Hown-Wen Chen
• Format: Others
• Language: zh-TW
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/81845200224257877230

碩士 === 大葉大學 === 資訊管理學系碩士班 === 95 === Traditional techniques of social network analysis mainly focused on identify the central members in a network, but rarely explored the interaction and relationships be-tween actors and sub-groups. A lot of information within a network is hence ignored. Therefore, a novel directional link analysis technique is proposed here. Instead, both dynamic and static relationships between actors and sub-groups will be investigated from the latent structure of a network. Enron e-mail corpus and indictment name list were used to illustrate and demon-strate the proposed techniques. First, the e-mail contact networks were classified by the topics of e-mail contents between July and Nov. 2001 which were regarded as the pe-riod of Enron fraud crisis. Following that, strongly connected components, communi-cation bridges between components and other topological information of the e-mail networks were found by the proposed Directional Link Analysis (DLA). To identify the employee involving fraud within the node set found by DLA. The experimental results illustrated the effectiveness and efficiency of the proposed methods by the average de-tection rate and the average running time being 83.07% and 1.47 seconds, respectively. Therefore, DLA and DNSA are useful as novel tools of automatic link analysis because of their efficiency and objectiveness.