Scalable Bayesian Inference for Time Series via Divide-and-Conquer

Abstract

Bayesian computational algorithms tend to scale poorly as data size increases. This had led to the de- velopment of divide-and-conquer-based approaches for scalable inference. These divide the data into subsets, perform inference for each subset in parallel, and then combine these inferences. While appeal- ing theoretical properties and practical performance have been demonstrated for independent observa- tions, scalable inference for dependent data remains challenging. In this work, we study the problem of Bayesian inference from very long time series. The literature in this area focuses mainly on approximate approaches that lack any rigorous theoretical guarantees and may provide arbitrarily poor accuracy in practice. We propose a simple and scalable divide-and-conquer method, and provide accuracy guaran- tees. Numerical simulations and real data applications demonstrate the effectiveness of our approach

Advisor(s)

David DUnson