Optimization of the Neural-Network-Based Multiple Classifiers Intrusion Detection System

In this paper, according to the difference between the attack categories, we adjust the 41-dimensional input features of the neural-network-based multiple classifiers intrusion detection system. After repeated experiment, we find that the every adjusted sub-classifier is better in convergence precision, shorter in training time than the 41-features sub-classifier, moreover, the whole intrusion detection system is higher in the detection rate, and less in the false negative rate than the 41-features multiple classifiers intrusion detection system. So, the scheme of the adjusting input features is able to optimize the neural-network-based multiple classifiers intrusion detection system, and proved to be feasible in practice.