Efficient Data Communication in SIoT: Hybrid Channel Attention Recurrent Transformer-Based Adaptive Marine Predator Algorithm for Reduced Energy Consumption

Abstract

The rapid development of technologies has attracted significant attention, with the social web and big data becoming key drivers of modern innovation. Although big data in the Social Internet of Things presents various energy-saving merits, problems such as network congestion and data communication reliability occur. In this article, a hybrid channel attention recurrent transformer-based adaptive marine predator algorithm is introduced to solve these problems. The main purpose of this approach is to improve the robustness and performance of SIoT systems. The hybrid channel attention recurrent transformer-based adaptive marine predator algorithm combines a hybrid recurrent neural network, a channel attention mechanism, and a transformer classifier. In this work, four datasets, including the water treatment plant, GPS trajectories, hepatitis dataset, and Twitter for sentiment analysis in Arabic are employed in validating the performance of a proposed model. The Savitzky–Golay filter is applied to reduce noise and eliminate unnecessary or irrelevant data. After data pre-processing, the hybrid channel attention recurrent transformer-based adaptive marine predator was introduced for classification, and this model is fine-tuned by the adaptive marine predator algorithm. In addition, the proposed model demonstrates strong scalability and applicability in real-world applications, making it an ideal solution for future Social Internet of Things systems.