Research Article
Anomaly Detection for Massive Data of Network Transmission Time Series Based on Graph Neural Network
@INPROCEEDINGS{10.1007/978-3-032-00303-4_3, author={Xingqi Gao and Guikun Cao and Yuren Liu and Zhenguo Wu}, title={Anomaly Detection for Massive Data of Network Transmission Time Series Based on Graph Neural Network}, proceedings={Advanced Hybrid Information Processing. 8th International Conference, ADHIP 2024, Jiaxing, China, September 20--22, 2024, Proceedings, Part III}, proceedings_a={ADHIP}, year={2026}, month={11}, keywords={Graph Neural Network Time Massive Data Series Network Transmission Anomaly Detection}, doi={10.1007/978-3-032-00303-4_3} }- Xingqi Gao
Guikun Cao
Yuren Liu
Zhenguo Wu
Year: 2026
Anomaly Detection for Massive Data of Network Transmission Time Series Based on Graph Neural Network
ADHIP
Springer
DOI: 10.1007/978-3-032-00303-4_3
Abstract
Most of the time series anomaly detection methods are unable to deeply explore the potential correlation between the data, and at the same time, there is an imbalance between the performance efficiency and the training optimization, and they also lack the interpretability of anomalies. Propose an unsupervised multivariate time-series data anomaly detection method based on graph neural networks and construct an anomaly detection model. The massive data of network transmission time series are preprocessed and the parameters are optimized. The model is jointly trained to complete the anomaly detection of massive data. The multi dataset validation of the proposed graph neural network anomaly detection method is effective, improving the experience of timing anomaly detection in industrial control systems and enhancing infrastructure security.