Genetic algorithm with Different Feature Selection Method for Intrusion Detection

Intrusion detection is used to protect the system from inside and outside attacks. Evolutionary algorithm has an important role in intrusion detection. Evolutionary algorithms are highly responsive for feature space reduction. The minimal number of features can improve the performance of an intrusion detection system. Thus we propose an intrusion detection system with various feature selection methods like information gain, mutual correlation, and cardinality of features. Genetic algorithm is applied into variable feature subset. The result depict that information based feature selection method can improve the detection rate. Accuracy of 87.54% is obtained in this model.