Improved Genetic Algorithm for Intrusion Detection System

Intrusion detection is one of the important security constraints for maintaining the integrity of information. Various approaches have been applied in past that are less effective to curb the menace of intrusion. The purpose of this paper is to provide an intrusion detection system (IDS), by modifying the genetic algorithm to network intrusion detection system. As we have applied attribute subset reduction on the basis of Information gain. So the training time and complexity reduced considerably. Moreover, we embedded a soft computing approach in rule generation makes the rule more efficient than hard computing approach used in existing genetic algorithm. Generated rule can detect attack with more efficiency. This model was verified using KDD'99 data set. Empirical result clearly shows the higher detection rates and low false positive rates.