Benchmarking Graph Representations and Graph Neural Networks for Multivariate Time Series Classification
Wennuo Yang, Shiling Wu, Yuzhi Zhou, Cheng Luo, Xilin He, Weicheng Xie, Linlin Shen, Siyang Song
TL;DR
This paper introduces a standardized benchmarking framework for graph-based multivariate time series classification (MTSC), evaluating 60 variants across three axes: node-feature definitions, edge-feature learning strategies, and Graph Neural Network (GNN) architectures, on 26 UEA MTSC datasets. It finds that node features strongly influence performance, with raw series often outperforming frequency-domain features, and that adaptive, especially multi-dimensional edge features, provide robust improvements across datasets. The MEGAT model, which uses multi-dimensional edge features, consistently outperforms other GNNs, underscoring the value of richer inter-variable relationship modeling. The authors provide open-source code to enable reproducible benchmarking and practical guidance for selecting node and edge representations in graph-based MTSC.
Abstract
Multivariate Time Series Classification (MTSC) enables the analysis if complex temporal data, and thus serves as a cornerstone in various real-world applications, ranging from healthcare to finance. Since the relationship among variables in MTS usually contain crucial cues, a large number of graph-based MTSC approaches have been proposed, as the graph topology and edges can explicitly represent relationships among variables (channels), where not only various MTS graph representation learning strategies but also different Graph Neural Networks (GNNs) have been explored. Despite such progresses, there is no comprehensive study that fairly benchmarks and investigates the performances of existing widely-used graph representation learning strategies/GNN classifiers in the application of different MTSC tasks. In this paper, we present the first benchmark which systematically investigates the effectiveness of the widely-used three node feature definition strategies, four edge feature learning strategies and five GNN architecture, resulting in 60 different variants for graph-based MTSC. These variants are developed and evaluated with a standardized data pipeline and training/validation/testing strategy on 26 widely-used suspensor MTSC datasets. Our experiments highlight that node features significantly influence MTSC performance, while the visualization of edge features illustrates why adaptive edge learning outperforms other edge feature learning methods. The code of the proposed benchmark is publicly available at \url{https://github.com/CVI-yangwn/Benchmark-GNN-for-Multivariate-Time-Series-Classification}.