Comparative analysis of CatBoost, LightGBM, XGBoost, RF, and DT methods optimised with PSO to estimate the number of k-barriers for intrusion detection in wireless sensor networks

Abstract The protection of borders is a critical concern for all countries, and Wireless Sensor Networks (WSNs) play a crucial role in assuring security by enabling intrusion detection and surveillance at border regions. This study presents an effective machine learning model designed to predict the number of k-barriers for rapid and robust intrusion detection and prevention in a rectangular area utilizing features extracted from a WSN through Monte-Carlo simulation. The proposed model is implemented using the CatBoost method, which is optimized with the metaheuristic Particle Swarm Optimisation (PSO) algorithm. To ensure a fair assessment, the performance of the proposed model is compared with state-of-the-art methods such as LightGBM, XGBoost, Random Forest (RF), and Decision Tree (DT) using metrics like MAE, MSE, RMSE, and R 2 . The proposed model demonstrates clear superiority, outperforming XGBoost, LightGBM, RF, and DT with an impressive R 2 value of 0.9998, MAE value of 0.6298, MSE value of 0.6018, and RMSE value of 0.7758. Additionally, the proposed model is evaluated against other existing methods that use the same dataset, and it produces the most favourable outcomes. Furthermore, the feature importance analysis reveals the network area and transmission range as the most influential features in predicting the number of k-barriers. In conclusion, the proposed CatBoost-PSO model exhibits exceptional performance and prediction accuracy, making it a promising approach for intrusion detection and prevention in WSN applications.