Intrusion Detection System for Anomaly in IoT Using Nonlinear Convergence Factor Using Salp Swarm Algorithm

Intrusion Detection Systems (IDS) for anomaly detection in Internet of Things (IoT) monitor network traffic and device behaviour to identify unusual patterns which denotes security threats. It helps protect IoT networks by detecting deviations from normal operations, potentially preventing cyberattacks. However, the propagation of IoT devices has made networks increasingly vulnerable to cyber threats, posing challenges in identifying anomalous behaviours within large-scale. Introducing a significant intrusion detection system for anomaly detection in IoT environments is critical to addressing evolving security threats and ensuring reliable performance. To overcome these challenges, the proposed Nonlinear Convergence Factor- Salp Swarm Algorithm (NCF-SSA) is suggested for anomaly detection at large scales, and to address security threats in IoT systems. Data pre-processing is carried out using hash encoding and normalization techniques. The NCF-SSA performs effectively with relevant feature selection from the input data, while Long Short-Term Memory (LSTM) performs classification. The proposed method attains a precision of 94.56% and an F1-score of 92.78% on the UNSW NB15 dataset.