Abstract:

The rapid development of wireless technologies and the growing dependency on WLAN infrastructures have significantly increased the surface of potential cyber threats. Among these, Evil Twin and Rogue Access Point attacks remain some of the most persistent and dangerous due to their ability to mimic legitimate network entities and deceive end users. This paper presents a comparative analysis of machine learning algorithms applied to the detection, prediction, and prevention of WLAN attacks, with a particular focus on the Evil Twin attack family. The study provides an overview of datasets used in research—namely AWID3 and NSL-KDD—and summarizes recent approaches identified through a literature review of IEEE publications from the past three years. The results highlight the consistently high accuracy of various supervised learning algorithms and emphasize the need for lightweight, computationally efficient models suitable for real-time wireless intrusion detection systems. The paper concludes with recommendations for future research directions, focusing on optimizing traditional machine learning algorithms for resource-constrained environments.