Feature-based Characterization and Clustering of Social Networks

• Main Authors: Chih-ShengLin, 林致聖
• Other Authors: Hui-Tang Lin
• Format: Others
• Language: en_US
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/2b68cg

碩士 === 國立成功大學 === 電腦與通信工程研究所 === 103 === How do we distinguish one type of social networks from another if topologies are the only available information? In this thesis, we proposed an approach to characterize social networks with techniques widely used in the social network analysis. These features are computed only based on the given topologies. With the aid of proposed characteristics, classification can then be performed between different types of networks. Our experiments show that a high accuracy can be achieved based on the proposed method. The approach can be used for advertisement distribution system, recommendation systems, and DGA-based botnet detection systems. Experiment also shows that the proposed system can be applied to anomaly detection system.