Forecasting Time Series with VARMA Recursions on Graphs

Abstract

Graph-based techniques emerged as a choice to deal with the dimensionality issues in modeling multivariate time series. However, there is yet no complete understanding of how the underlying structure could be exploited to ease this task. This work provides contributions in this direction by considering the forecasting of a process evolving over a graph. We make use of the (approximate) time-vertex stationarity assumption, i.e., timevarying graph signals whose first and second order statistical moments are invariant over time and correlated to a known graph topology. The latter is combined with VAR and VARMA models to tackle the dimensionality issues present in predicting the temporal evolution of multivariate time series. We find out that by projecting the data to the graph spectral domain: (i) the multivariate model estimation reduces to that of fitting a number of uncorrelated univariate ARMA models and (ii) an optimal low-rank data representation can be exploited so as to further reduce the estimation costs. In the case that the multivariate process can be observed at a subset of nodes, the proposed models extend naturally to Kalman filtering on graphs allowing for optimal tracking. Numerical experiments with both synthetic and real data validate the proposed approach and highlight its benefits over state-of-the-art alternatives.

submitted to the IEEE Transactions on Signal Processing

References in corpus (2)

• Discrete Signal Processing on Graphs
• Regularized Tyler's Scatter Estimator: Existence, Uniqueness, and Algorithms

Cited by in corpus (14)

• Graph Signal Processing: History, Development, Impact, and Outlook
• Reconstruction of Time-varying Graph Signals via Sobolev Smoothness
• Graph Deep Learning for Time Series Forecasting
• A Robust and Efficient Multi-Scale Seasonal-Trend Decomposition
• Joint Estimation and Prediction of City-wide Delivery Demand: A Large Language Model Empowered Graph-based Learning Approach
• Robust Graph Filter Identification and Graph Denoising from Signal Observations
• Joint Time-Vertex Fractional Fourier Transform
• Preconditioned Gradient Descent Algorithm for Inverse Filtering on Spatially Distributed Networks
• Polynomial graph filter of multiple shifts and distributed implementation of inverse filtering
• Spatiotemporal Covariance Neural Networks
• Graph Neural Networks for Distributed Linear-Quadratic Control
• Methods of Adaptive Signal Processing on Graphs Using Vertex-Time Autoregressive Models
• Online Distributed Learning over Graphs with Multitask Graph-Filter Models
• Stationarity of Time-Series on Graph via Bivariate Translation Invariance