Anomaly-based intrusion detection system using Harris Hawks optimisation with a sigmoid neuron network

This study introduces an innovative approach, merging Harris Hawks optimisation (HHO) with a sigmoid neuron network (SN), to enhance anomaly-based intrusion detection systems (ADS) performance. The resultant SN-HHO hybrid model aims to elevate detection rates and lower false positive rates (FPRs) within ADS. Evaluation across five datasets - UNSW-NB15, CIDDS-001, NSL-KDD, AWID3, and CICDDoS2019 - reveals heightened accuracy and faster convergence compared to existing methods. This work underscores the potential synergy of meta-heuristic optimisation and artificial neural networks, offering a promising strategy to fortify IDS performance and reliability, thus presenting a novel direction for advancing anomaly detection practices.