Graph Attention Network and Informer for Multivariate Time Series Anomaly Detection

Mengmeng Zhao,Haipeng Peng,Lixiang Li,Yeqing Ren

DOI: https://doi.org/10.3390/s24051522

IF: 3.9

2024-02-27

Sensors

Abstract:Time series anomaly detection is very important to ensure the security of industrial control systems (ICSs). Many algorithms have performed well in anomaly detection. However, the performance of most of these algorithms decreases sharply with the increase in feature dimension. This paper proposes an anomaly detection scheme based on Graph Attention Network (GAT) and Informer. GAT learns sequential characteristics effectively, and Informer performs excellently in long time series prediction. In addition, long-time forecasting loss and short-time forecasting loss are used to detect multivariate time series anomalies. Short-time forecasting is used to predict the next time value, and long-time forecasting is employed to assist the short-time prediction. We conduct a large number of experiments on industrial control system datasets SWaT and WADI. Compared with most advanced methods, we achieve competitive results, especially on higher-dimensional datasets. Moreover, the proposed method can accurately locate anomalies and realize interpretability.

engineering, electrical & electronic,chemistry, analytical,instruments & instrumentation

What problem does this paper attempt to address?

The problem that this paper attempts to solve is that in industrial control systems (ICSS), as the feature dimension increases, the performance of most existing anomaly detection algorithms drops sharply. Specifically, the paper proposes a multivariate time - series anomaly detection scheme based on Graph Attention Network (GAT) and Informer. This method aims to improve the accuracy of anomaly detection by combining the effective sequence feature - learning ability of GAT and the excellent performance of Informer in long - time - series prediction, especially on high - dimensional data sets. In addition, the paper also detects anomalies in multivariate time series by using long - term prediction loss and short - term prediction loss to achieve more accurate anomaly location and interpretation capabilities.

The main contributions of the paper include:
1. Proposing a new time - series anomaly detection method based on graph attention and Informer, which is based on time - series prediction.
2. Using graph attention and GRU to learn the short - term features of time series, using Informer to learn the long - term features of time series, and then guiding short - term time - series prediction through long - term time - series prediction to complete the final anomaly judgment.
3. In order to improve the accuracy of the model, short - term prediction loss and long - term prediction loss are considered.
4. Experiments on the SWaT and WADI data sets show that the proposed model has higher time - series anomaly detection performance.

Through these methods, the paper aims to provide a more effective and accurate multivariate time - series anomaly detection solution, especially when dealing with high - dimensional data.