A Long Short-Term Memory Learning Based Malicious Node Detection for Clustering in Wireless Sensor Networks

WSNs (wireless sensor networks) are essential for everyday tasks such as security, medical care, and industrial automation, making them susceptible to many kinds of Denial of service (DoS) attacks. This work explores the various models such as Artificial neural-networks (ANN), Recurrent neural-networks (RNN), and Long Short-Term Memory (LSTM) networks for attack detection. The models’ performance is analyzed across multiple evaluation metrics, including RMSE, R2 score, and accuracy. The results show that LSTM outperform other models with overall RMSE 0.17, R2 score 0.9033 and accuracy 99.08%. The findings showcase the potential of various models to improve WSN security through accurate attack identification and offer insights into model selection for specific applications.