Graph-CNN Hybrid Advance Model for Accurate Anomaly Detection in Multivariate Time Series IoT Streams

Abstract The Internet of Things (IoT) has revolutionized various industries by enabling the collection of vast amounts of data from interconnected devices. However, the sheer volume and complexity of IoT data pose significant challenges for anomaly detection, especially in multivariate time series data. Traditional methods struggle to effectively capture the intricate relationships and temporal dependencies within such data. In response to these challenges, this paper introduces a pioneering approach that harnesses the combined power of Graph Neural Networks (GNNs) and Convolutional Neural Networks (CNNs) for anomaly detection in multivariate time series IoT data. The novelty of our approach lies in its holistic integration of GNNs and CNNs, enabling the comprehensive analysis of both global dependencies and local temporal features within the IoT data. Through rigorous experimentation on real-world datasets, showcases the efficacy and superiority of our method over existing state-of-the-art techniques. Our approach consistently achieves superior performance metrics, compared to state-of-the-art methods. Underscoring its effectiveness in accurately detecting anomalies amidst the complexities of IoT data.