A Machine Learning-based Wireless LAN Defence Index Detection System

Abstract

Denial of service attack, one of the biggest network attacks that harm wireless networks, has the purpose of interfering with or destroying communication by consuming network bandwidth resources and server memory resources. Aiming at the current wireless LAN attack index detection method exists problems such as poor detection accuracy, this paper proposes a wireless LAN defence index detection method based on an improved heuristic optimization algorithm optimised neural network. Initially, the classification of WLAN attacks and their corresponding feature sets were examined to devise a mapping model specifically for the detection of WLAN defence indices. Utilizing the kernel principal component analysis (KPCA) technique, key features of WLAN attacks were identified and extracted. Subsequently, the refined Tuna Swarm Optimization (TSO) algorithm was integrated to enhance the neural network's parameters, leading to the development of a BP neural network-based WLAN defence index detection model that is underpinned by the optimized TSO algorithm. Ultimately, the efficacy of the detection model was scrutinized through simulated WLAN attack scenarios, which confirmed the proposed method's elevated precision. The experimental findings indicate that this approach has successfully enhanced the detection system's accuracy and addressed the issue of feature set redundancy within WLAN attacks.