Comparative Analysis of Machine Learning Algorithms for Pulse Jammer Detection

This work compares various machine learning algorithms for predicting the presence of a random on/off jammer in wireless communication systems. We analyze four machine learning classification algorithms, Decision Trees, Support Vector Machines, K-nearest neighbors, and Random Forest. The study applies data pre-processing techniques to improve detection performance and assesses the algorithms using metrics such as accuracy, precision, recall, and F1 score. Among the tested methods, Random Forest and Decision Trees achieved the highest performance with an accuracy of 97.5%, precision of 1.00, recall of 0.95, and an $F 1$ score of 0.97. SVM demonstrated strong precision at 1.00 but slightly lower accuracy at $\mathbf{9 2. 5 \%}$, while KNN exhibited the lowest accuracy at 85%. These results highlight the effectiveness of Random Forest and Decision Trees for robust and efficient jammer detection. The findings have significant implications for enhancing the resilience of wireless communication systems against jamming attacks. In particular, we show the use-case of our analytical work in the classification-based application of wireless communication systems, specifically jammer detection.