A Neural Network Fusion Feature Selection for Network Intrusion Detection

Abstract

Network Intrusion Detection often utilized to shield computer networks from a multitude of attacks, poses a significant challenge in the realm of system and network security due to the pervasive nature of network threats. At present, Deep Learning methods of Neural Networks are often used in Network Intrusion Detection, but the performance of classifiers may vary from data set to data set. The main reason for this is that there are some errors or duplicate features. To solve this problem, three feature selection methods are combined with Neural Networks to improve the performance of Network Intrusion Detection. In the experimental evaluation phase, this paper uses experimental results demonstrate the superiority of the proposed method in comparison to existing Network Intrusion Detection techniques, by identifying significant and intimately related features as tested on the KDD99 dataset for feasibility.