Enhancing cybersecurity in IoT networks: SLSTM-WCO algorithm for anomaly detection

Abstract

Internet of Things (IoT) security refers to different aspects of security, including methods, tactics, and technologies used to protect these devices from unauthorized access. However, it is connected with multiple physical devices to perform huge tasks simultaneously and secure the data transmitted through the IoT network. Furthermore, the IoT is used to transmit sensitive data and validate security performance. Mostly Machine Learning (ML) algorithms are widely utilized for the process of anomaly detection. However the application of the ML model fails to detect attacks in IoT, to overcome this, a novel Stacked Long Short Term Memory based Willow Catkin Optimization (SLSTM-WCO) algorithm is proposed to detect intrusion anomalies in IoT networks. The complex patterns and abnormalities are predicted by determining the regularization method. Also, the deep learning (DL) model such as stacked LSTM detects the anomaly accurately and improves the effectiveness. The detection performance is validated by using benchmark datasets such as BoT-IoT, IoT network Intrusion, IoT-23, MQTT, and MQTTset which enhanced the efficiency. The outcome of the SLSTM-WCO method improved accuracy by 99.49% and improved anomaly detection in IoT networks compared to existing methods.