Using Genetic Algorithm to Supporting Artificial Neural Network for Intrusion Detection System

Due to the rece nt trend of technologies to use the network based systems, detecting them from threats become a crucial issue. This paper investigates applying the following methods to detect the attacks in network: Genetic Algorithms (GA) with Artificial Neural Networks classifier, Modified Mutual Information Feature selection (MMlFS), Linear Correlation Feature Selection (LCFS), and Forward Feature Selection (FFS). Here, the capability of feature selection of LCFS, FFSA, MMIFS, and GA-ANN has been explored. We use KDD CPU dataset to obtain the results; which shows around 99% accuracy of applied methods in detecting threads. The requirement of GA with ANN is 18 features and there is respectively a requirement of 24, 21, and 31 features for MMIFS, LCFS and FFSA for efficiently detecting the attacks.