TOPSIS based multi-criteria decision making of feature selection techniques for network traffic dataset

R. Singh, Harish Kumar, R.K. Singla

Research output: Contribution to journalArticlepeer-review

14 Citations (Scopus)

Abstract

Intrusion detection systems (IDS) have to process millions of packets with many features, which delay the detection of anomalies. Sampling and feature selection may be used to reduce computation time and hence minimizing intrusion detection time. This paper aims to suggest some feature selection algorithm on the basis of The Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS). TOPSIS is used to suggest one or more choice (s) among some alternatives, having many attributes. Total ten feature selection techniques have been used for the analysis of KDD network dataset. Three classifiers namely Naïve Bayes, J48 and PART have been considered for this experiment using Weka data mining tool. Ranking of the techniques using TOPSIS have been calculated by using MATLAB as a tool. Out of these techniques Filtered Subset Evaluation has been found suitable for intrusion detection in terms of very less computational time with acceptable accuracy.
Original languageEnglish
Pages (from-to)4598-4604
Number of pages7
JournalInternational Journal of Engineering and Technology
Volume5
Issue number6
Publication statusPublished - 31 Dec 2013

Fingerprint

Dive into the research topics of 'TOPSIS based multi-criteria decision making of feature selection techniques for network traffic dataset'. Together they form a unique fingerprint.

Cite this