Rough Set-hypergraph-based Feature Selection Approach for Intrusion Detection Systems

  • M.R. Gauthama Raman School of Computing, SASTRA University, Thanjavur, India
  • K. Kannan Department of Mathematics, SASTRA University, Thanjavur, India
  • S.K. Pal Scientific Analysis Group, New Delhi - 110 054, India
  • V. S. Shankar Sriram School of Computing, SASTRA University, Thanjavur, India
Keywords: Intrusion detection systems, rough set theory, hyper graph, feature selection, ?-Helly property

Abstract

Immense growth in network-based services had resulted in the upsurge of internet users, security threats and cyber-attacks. Intrusion detection systems (IDSs) have become an essential component of any network architecture, in order to secure an IT infrastructure from the malicious activities of the intruders. An efficient IDS should be able to detect, identify and track the malicious attempts made by the intruders. With many IDSs available in the literature, the most common challenge due to voluminous network traffic patterns is the curse of dimensionality. This scenario emphasizes the importance of feature selection algorithm, which can identify the relevant features and ignore the rest without any information loss. In this paper, a novel rough set κ-Helly property technique (RSKHT) feature selection algorithm had been proposed to identify the key features for network IDSs. Experiments carried using benchmark KDD cup 1999 dataset were found to be promising, when compared with the existing feature selection algorithms with respect to reduct size, classifier’s performance and time complexity. RSKHT was found to be computationally attractive and flexible for massive datasets.

Author Biographies

M.R. Gauthama Raman, School of Computing, SASTRA University, Thanjavur, India

Mr M.R. Gauthama Raman obtained his BE from Anna University, Chennai, in 2012 and MTech from B.S. Abdur Rahman University, Chennai, in 2014. Currently pursuing his PhD at School of Computing, SASTRA University, Thanjavur.
K. Kannan, Department of Mathematics, SASTRA University, Thanjavur, India

Mr Krithivasan Kannan is a Professor in the Department of Mathematics, SARTRA University, Thanjavur, INDIA. He obtained his Bachelor’s and Master’s degrees from the University of Madras, India, in 1980 and 1982, respectively. He was conferred PhD in Mathematics in the area of Computational Fluid Dynamics by Alagappa University, Karaikudi, India, in 2000. His specific areas of interest include : Combinatorial optimisation, artificial neural networks and hyper graph-based image processing.
S.K. Pal, Scientific Analysis Group, New Delhi - 110 054, India
Dr S.K. Pal did his post-graduation in Computer Science from the J.K. Institute of Applied Physics, Electronics & Communications, University of Allahabad in 1990 and PhD from University of Delhi in the area of Information Security. Presently working as Scientist G, Scientific Analysis Group, Delhi. His research interests include: Digital signal processing, cryptology, multimedia and network security, information hiding and soft computing
V. S. Shankar Sriram, School of Computing, SASTRA University, Thanjavur, India
Dr Shankar Sriram V.S. is an Associate Professor in School of Computing, SASTRA University, Thanjavur, Tamil Nadu. He received his BSc from Madurai Kamaraj University, Madurai, India in 1997. He obtained his MCA from Madurai Kamraj University, Madurai, in 2000. He received his PhD from Birla Institute of Technology, Mesra, in 2010.
Published
2016-10-31
How to Cite
Raman, M., Kannan, K., Pal, S., & Sriram, V. S. (2016). Rough Set-hypergraph-based Feature Selection Approach for Intrusion Detection Systems. Defence Science Journal, 66(6), 612-617. https://doi.org/10.14429/dsj.66.10802
Section
Special Issue Papers