Emerging Intuitionistic Fuzzy Classifiers for Intrusion Detection System

One of the toughest challenges in Intrusion Detection System is uncertainty handling. he normal and the abnormal behaviors in networked computers are hard to predict as the boundaries cannot be well defined. The prediction of the normal or abnormal behaviors is done by the comparison with predefined classes to find the most similar one. This prediction process may generate false alarms in many anomaly based intrusion detection systems. Consequently, we observed uncertainty where there is a fair chance of the existence of a non-null hesitation part at each moment of evaluation of an unknown object. A new technique is implemented in this paper using Intuitionistic fuzzy logic which is a generalization of fuzzy logic. In this model the false alarm rate in determining intrusive activities can be reduced. A set of Intuitionistic fuzzy rules can be used to define the normal, abnormal and indeterministic behavior in a computer network. An Intuitionistic fuzzy inference algorithm can be applied over such rules to determine when an intrusion is in progress. The main problem with this approach is to generate good Intuitionistic fuzzy classifiers to detect intrusions. The rules generated by Intuitionistic fuzzy classifiers are fine tuned using improvised genetic algorithm that can detect anomalies and some specific intrusions. The main idea is to evolve three rules, one for the normal class, second for the abnormal class and third of indeterministic class using KDD Cup 99 Dataset. This paper exhibits the performance of Emergent Intuitionistic fuzzy classifiers in intrusion detection.