Identifying Important Features for Intrusion Detection Using Support Vector Machines and Neural Networks

  • Authors:
  • Andrew H. Sung;Srinivas Mukkamala

  • Affiliations:
  • -;-

  • Venue:
  • SAINT '03 Proceedings of the 2003 Symposium on Applications and the Internet
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

Intrusion detection is a critical component ofsecure information systems. This paper addresses theissue of identifying important input features in buildingan intrusion detection system (IDS). Since eliminationof the insignificant and/or useless inputs leads to asimplification of the problem, faster and more accuratedetection may result. Feature ranking and selection,therefore, is an important issue in intrusion detection.In this paper we apply the technique of deletingone feature at time to perform experiments on SVMsand neural networks to rank the importance of inputfeatures for the DARPA collected intrusion data.Important features for each of the 5 classes of intrusionpatterns in the DARPA data are identified.It is shown that SVM-based and neural networkbased IDSs using a reduced number of features candeliver enhanced or comparable performance. An IDSfor class-specific detection based on five SVMs isproposed.