A Study of Joint Graph Inference and Forecasting
This work addresses forecasting in multivariate time series for researchers, but it is incremental as it builds on existing models without introducing a new paradigm.
The study compared four recent models that use graph neural networks (GNNs) for multivariate time series forecasting, focusing on latent graph inference, and proposed novel combinations of these architectures based on ablation analyses.
We study a recent class of models which uses graph neural networks (GNNs) to improve forecasting in multivariate time series. The core assumption behind these models is that there is a latent graph between the time series (nodes) that governs the evolution of the multivariate time series. By parameterizing a graph in a differentiable way, the models aim to improve forecasting quality. We compare four recent models of this class on the forecasting task. Further, we perform ablations to study their behavior under changing conditions, e.g., when disabling the graph-learning modules and providing the ground-truth relations instead. Based on our findings, we propose novel ways of combining the existing architectures.