Intrusion Detection Using Fuzzy Stochastic Local Search Classifier

Abstract

This paper proposes a stochastic local search classifier combined with the fuzzy logic concepts for intrusion detection. The proposed classifier works on knowledge base modeled as a fuzzy rule "if-then" and improved by using a stochastic local search. The method is tested on the Benchmark KDD'99 intrusion dataset and compared with other existing techniques for intrusion detection. The results are encouraging and demonstrate the benefit of the proposed approach.