Bayesian forecasting of multivariate time series: Scalability, structure uncertainty and decisions

TL;DR

This paper reviews recent advances in Bayesian state-space models for multivariate time series, emphasizing scalability, structure uncertainty, and decision-making across various large-scale applications.

Contribution

It introduces the decouple/recouple concept enabling scalable Bayesian modeling for large multivariate time series data in diverse domains.

Findings

Effective application of state-space models to large-scale data

Use of dynamic graphical models for forecasting and volatility analysis

Multi-scale approaches for discrete time series forecasting

Abstract

I overview recent research advances in Bayesian state-space modeling of multivariate time series. A main focus is on the decouple/recouple concept that enables application of state-space models to increasingly large-scale data, applying to continuous or discrete time series outcomes. The scope includes large-scale dynamic graphical models for forecasting and multivariate volatility analysis in areas such as economics and finance, multi-scale approaches for forecasting discrete/count time series in areas such as commercial sales and demand forecasting, and dynamic network flow models for areas including internet traffic monitoring. In applications, explicit forecasting, monitoring and decision goals are paramount and should factor into model assessment and comparison, a perspective that is highlighted.