Social Network Security Using Anomaly Detection
Mohammad Arshi Saloot, S. Durga Bhavani · ASME Press eBooks · 2012
The paper describes a novel way to provide security for social network sites. The proposed system constructed from sophisticated architecture and methods which presented in Data mining context. The paper illustrates how to capture and analysis relation between participants. In addition, messages have been parsed and the semantic of message's topic has been discovered. An unsupervised anomaly detection algorithm is applied to real life email dataset. Eventually, experimental results disambiguate outlier messages and clarify nodes relations. Combination of social network analyzing and messages anomaly detection produces an efficient way to design and implementing secure social network.