Research on Anomaly Detection and Prediction of Internet Medical Network Traffic Based on Graph Neural Networks and Temporal Convolution

Abstract

With the rapid development of internet-based healthcare, ensuring the stable operation of its network is crucial, and anomaly detection and prediction of network traffic play a key role in this regard. This study provides a systematic review of the application of graph neural networks (GNNs) and temporal convolution in the field of internet medical network traffic. GNNs can effectively capture the complex relational features among nodes in medical network traffic data, accurately extracting spatial information and assisting in analyzing traffic correlations between different healthcare service nodes. Temporal convolution, on the other hand, excels at handling time-series data, enabling a deep analysis of the dynamic trends in internet medical network traffic, such as fluctuations in traffic patterns at different time intervals. By integrating these two approaches, a more powerful model can be built to comprehensively analyze internet medical network traffic from a spatiotemporal perspective. This hybrid approach not only enhances the precision of anomaly detection, effectively identifying various abnormal traffic patterns, but also improves the reliability of network traffic forecasting. This theoretical study assists internet medical network administrators in timely detection and resolution of network anomalies, providing a scientific basis for the rational allocation of network resources. It holds significant theoretical and practical value in promoting the security and efficient operation of internet-based healthcare networks.