A Note on Implementing Recurrence Quantification Analysis for Network Anomaly Detection

  • Ch. Aswani Kumar Vellore Institute of Technology University, Vellore
  • Bhargavi K. Vellore Institute of Technology University, Vellore
  • Garima Jalota Vellore Institute of Technology University, Vellore
Keywords: Cross Recurrence Plots, Network anomaly detection, Non Linear Analysis, Recurrence Quantification Analysis, Support Vector Machines.

Abstract

This paper deal with the network anomaly detection, based on the analysis of non-stationary properties that occur in the aggregated IP traffic flows. We use recurrence quantification analysis (RQA), a mathematical nonlinear technique to achieve this task. The objective is to model the standard network traffic and report any deviation from it. We create a baseline from which we derive the RQA parameters. Using these parameters we explore the hidden recurrence patterns in the network traffic. Further, the detection is analysed using the support vector machine to classify the deviations from the regular traffic. Experiments are conducted on Vellore Institute of Technology University campus network traffic data to validate the model.

Defence Science Journal, 2012, 62(2), pp.112-116DOI:http://dx.doi.org/10.14429/dsj.62.1171

Author Biographies

Ch. Aswani Kumar, Vellore Institute of Technology University, Vellore
Dr Ch. Aswani Kumar received PhD (Computer Science) from Vellore Institute of Technology (VIT) University, India. Currently working as a Associate Professor at School of Information Technology and Engineering, VIT University, Vellore, India. He has published 35 refereed research papers so far in various national, international journals and conferences. His research interests include: Data mining, formal concept analysis, information security, and machine intelligence.
Bhargavi K., Vellore Institute of Technology University, Vellore
Ms Bhargavi is a BTech (IT) from VIT University, Vellore, India. Presently she is working as Database Administrator in Ford Technology Services India, Chennai. Her research interests include: Databases and networks.
Garima Jalota, Vellore Institute of Technology University, Vellore
Ms Garima is a BTech (IT) from VIT University, Vellore, India. Presently she is working as Associate Software Engineer in Robert Bosch Engineering & Business Solutions, Bangalore. Her research interests include: Database management systems and networks.
Published
2012-03-13
How to Cite
Kumar, C., K., B., & Jalota, G. (2012). A Note on Implementing Recurrence Quantification Analysis for Network Anomaly Detection. Defence Science Journal, 62(2), 112-116. https://doi.org/10.14429/dsj.62.1171
Section
Computers & Systems Studies