Wireless Communication Strategies to Achieve Secrecy and Covertness with Machine Learning Optimization

With the increasing dependence on wireless communication systems, the risk of cyberattacks has escalated, highlighting the urgent need for robust intrusion detection mechanisms to protect data integrity and confidentiality. This paper introduces a novel approach to bolster security in wireless communications through the application of machine learning optimization techniques. We assess a variety of algorithms, including Support Vector Machines (SVM), Logistic Regression, Naive Bayes, and Random Forest, with initial accuracies recorded at 48.5%, 87.5%, 81.5%, and 97.9%, respectively. To enhance detection capabilities further, we propose a hybrid model that integrates the strengths of Random Forest, XGBoost, and LightGBM, employing a voting strategy that yields an impressive overall accuracy of 98.3%. Additionally, we apply hyperparameter tuning methods, including grid search, to refine model performance and improve the accuracy of intrusion detection. The results underscore the critical role of machine learning techniques in cybersecurity, demonstrating the effectiveness of ensemble methods in achieving significant accuracy gains. This research not only advances the field of wireless communication security but also lays the ground work for future explorations at the intersection of cybersecurity and machine learning.