Title: A review of social network centric anomaly detection techniques

Authors: Ravneet Kaur; Sarbjeet Singh

Addresses: Department of Computer Science and Engineering, UIET, Panjab University, Chandigarh – 160014, India ' Department of Computer Science and Engineering, UIET, Panjab University, Chandigarh – 160014, India

Abstract: Online social networks have gained much attention in the recent years in terms of their analysis for usage as well as detection of abnormal activities. Anomalous activities arise when someone shows a different behaviour than others in the network. Presence of these anomalies may pose a number of problems which need to be addressed. This paper discusses different types of anomalies and their novel categorisation based on various factors. A review of various techniques used for detecting anomalies along with underlying assumptions and reasons for the presence of such anomalies is also covered. A special reference is made to different data mining approaches used to detect anomalies. However, the major focus of paper is the analysis of social network centric anomaly detection approaches which are broadly classified as behaviour-based, structure-based and spectral-based. Each one of this classification further incorporates a number of techniques which are discussed in the paper.

Keywords: classification; clustering; centrality; data mining; graph-based anomaly detection; online social networks; social network analysis; SNA; proximity; static networks; dynamic networks.

DOI: 10.1504/IJCNDS.2016.080582

International Journal of Communication Networks and Distributed Systems, 2016 Vol.17 No.4, pp.358 - 386

Accepted: 17 May 2016
Published online: 30 Nov 2016 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article