Applying Data Mining of Fuzzy Association Rules to Network Intrusion Detection

Abstract

This paper describes the use of fuzzy logic in the implementation of an intelligent intrusion detection system. The system uses a data miner that integrates Apriori and Kuok's algorithms to produce fuzzy logic rules that capture features of interest in network traffic. Using an inference engine, implemented using FuzzyJess, the intrusion detection system evaluates these rules and gives network administrators indications of the firing strength of the ruleset. The resulting system is capable of adapting to changes in attack signatures. In addition, by identifying relevant network traffic attributes, the system has the inherent ability to provide abstract views that support network security analysis. Examples and experimental results using intrusion detection datasets from MIT Lincoln Laboratory demonstrate the potential of the approach