Bayesian Computation in Dynamic Latent Factor Models

TL;DR

This paper introduces a novel copula-based Bayesian filtering method for dynamic multivariate models, significantly improving computational efficiency in nonlinear, non-Gaussian time series analysis.

Contribution

It develops a new copula construction integrated into multiscale dynamic models, enabling faster Bayesian filtering and forecasting for complex time series.

Findings

Dramatic speed-up in online Bayesian computations.

Comparable filtering and forecasting accuracy to existing methods.

Effective handling of heterogeneous non-negative count data.

Abstract

Bayesian computation for filtering and forecasting analysis is developed for a broad class of dynamic models. The ability to scale-up such analyses in non-Gaussian, nonlinear multivariate time series models is advanced through the introduction of a novel copula construction in sequential filtering of coupled sets of dynamic generalized linear models. The new copula approach is integrated into recently introduced multiscale models in which univariate time series are coupled via nonlinear forms involving dynamic latent factors representing cross-series relationships. The resulting methodology offers dramatic speed-up in online Bayesian computations for sequential filtering and forecasting in this broad, flexible class of multivariate models. Two examples in nonlinear models for very heterogeneous time series of non-negative counts demonstrate massive computational efficiencies relative to existing simulation-based methods, while defining similar filtering and forecasting outcomes.