Genetic fuzzy system for intrusion detection: Analysis of improving of multiclass classification accuracy using KDDCup-99 imbalance dataset

Genetic fuzzy systems (GFSs) hybridize the approximate reasoning method of fuzzy systems with the learning capability of evolutionary algorithms. The objective of this paper is to focus on an important class of problems in the field of network intrusion detection, namely, the class imbalance problem, one of the problems strongly tied with the classification of the database of intrusion detection. In this work we have used a fuzzy association rule based classification method, to obtain an accurate and compact fuzzy rule-based classifier with a low computational cost. We have proposed the use of a novel fitness function to deal with the problem of imbalance dataset in the genetic post processing phase for rule selection and parameter tuning. The efficiency of the proposed system has been shown through a complete detailed experimental comparative study with well-known classifiers reported in the IDS literature. Experiments were performed with KDD Cup 99 intrusion detection benchmark data set as an example of a network traffic data.