Chaotic Sparrow Search Algorithm with Deep Learning for Anomaly Detection in Internet of Things

Anomaly Detection (AD) systems play a crucial role in identifying potential cyber-attacks or data breaches by recognizing patterns of irregular data within the Internet of Things (IoT). Standard Machine Learning (ML) techniques often prove inefficient in the face of unpredictable network behaviors and diverse anomalous methods. Deep Learning (DL) has appeared as an efficient as well as robust approach for AD, capable of classifying abnormal behaviors or patterns in data. A Chaotic Sparrow Search Algorithm with DL utilizing Recurrent Neural Network (RNN-CSSA) is proposed for AD in IoT aided sustainable smart cities. A primary objective of RNN-CSSA technique is to accurately detect anomalies in IoT aided smart cities. To achieve this, the RNN-CSSA system employs Binary Pigeon optimization Algorithm (BPEO) for effective feature selection. Additionally, CSSA technique utilizes the RNN approach for classifying anomalies. The evaluations results of the RNN-CSSA algorithm utilized the benchmark databases such as UCI-SECOM and UNSW NB-15. The results exhibit an effectiveness of proposed RNN-CSSA methodology, achieving accuracies of 99.54% and 99.55% in UNSW NB-15 and UCISECOM, when compared with existing models.