Improving Accuracy of IDS Using Genetic Algorithm and Multilayer Perceptron Network

Adoption of the use of Internet and increased in reliance on technology, the security of computer network and information system become more and more important for all types of organizations. Since the security is all-important for any types of organizations, we make an empirical study on network intrusion detection system which is one of the essential layers of organization security. In this study, genetic algorithm and multilayer perceptron network are used as methodologies. We divide the dataset into three parts according to the protocol and then applied a genetic algorithm for attribute selection. Multilayer perceptron network is used to train for each classifier. We examine performance differences between some recent research work and our system. The result shows that our protocol-based Genetic Algorithm for Multilayer Perceptron Network (GA-MLP) model has slightly increased in detection rate.