Sin-Cos Fusion-Based African Vulture Optimization Algorithm for Feature Selection in Detecting DDoS Attack in IoT Networks

This paper presents a performance investigation of the Sin-Cos Fusion-Based Improved African Vulture Optimization Algorithm for detecting DDoS attacks in IoT networks. Various competing state-of-the-art methods were employed to compare the performances with SCIAVOA using three bench-mark datasets, namely NSL-KDD, TON-IoT, and UNSW. The evaluation metrics evident from the robust performance of SC-IAVOA. Compared across datasets, SC-IAVOA greatly outperforms for precision, with 98.68% for NSL-KDD, 99.40% for TON-IoT, and 99.43% for UNSW; for recall, 98.18% for NSL-KDD, 99.88% for TON-IoT, and 99.53% for UNSW; and for F-score, 98.68% for NSL-KDD, and 99.58% and 99.59% for TON-IoT and UNSW, respectively. These findings further underpin the efficacy of SC-IAVOA in feature optimization for accurate and reliable IoT DDoS attack detection systems, leading to a much greater improvement in network security.