Skip to content
DraconiticResearch
Search Sources API
203,203 studies31 sourcesUpdated 24 min ago
Peer-reviewedCC BYJournal of the American Statistical Association

Scalable Bayesian Inference for Time Series via Divide and Conquer

  • Rihui Ou1,2
  • Lachlan Astfalck3,4
  • Deborshee Sen1
  • David B. Dunson1
4 affiliations
  1. 1Department of Statistical Science, Duke University
  2. 2Duke University#TAB#
  3. 3School of Mathematics and Statistics, University of New South Wales
  4. 4School of Physics, Mathematics and Computation, The University of Western Australia
Open at sourceDOICite

Abstract

Bayesian computation often scales poorly with increasing data size, motivating developments such as divide-and-conquer approaches for scalable inference. These methods partition the data into subsets, perform parallel inference on each subset, and aggregate the results into a single posterior. Appealing theoretical properties and practical performance have been demonstrated for independent data; however, methods for dependent data remain challenging. Existing methods rely on ad hoc approximations with limited theoretical guarantees that lead to potentially poor accuracy in practice. Here, we focus on time-series data and propose a simple, scalable divide-and-conquer method for dependent time series, with theoretically rigorous accuracy guarantees. Simulation studies and real-data examples are used to empirically verify the effectiveness of our approach.

Related

  1. Spectral divide-and-conquer MCMC for long stationary time series
    Zixuan Wang, Matias Quiroz, Feng Li +2
    arXivPreprintarXiv licenceFull textStatisticsarXiv
  2. Two moments suffice for Bessel approximation: conditioned differences, contingency tables, and probabilistic divide-and-conquer
    Stephen DeSalvo
    arXivPreprintarXiv licenceFull textCombinatorics & discrete mathematicsProbabilityarXiv
  3. Divide-and-Conquer Neural Bayes Estimation for Spatial and Spatiotemporal Point Process Models
    Bokgyeong Kang, Andrew Zammit-Mangion, Erin M. Schliep +1
    arXivPreprintarXiv licenceFull textStatisticsarXiv
  4. Scaling Spatially Varying Distributed Lag Nonlinear Models for Nationwide Small-Area Data: A Divide-and-Conquer Approach
    Sara Rutten, Thomas Neyens, Elisa Duarte +1
    arXivPreprintCC BY-NC-ND 4.0Full textStatisticsarXiv
  5. Compressed Active Subspaces for Scalable Bayesian Inference
    Thomas Flynn, Sanket Jantre, Byung-Jun Yoon +1
    arXivPreprintCC BY 4.0Full textStatisticsarXiv

Cite

BibTeXRIS
BibTeX
@article{ou2026scalable,
  title = {{Scalable Bayesian Inference for Time Series via Divide and Conquer}},
  author = {Rihui Ou and Lachlan Astfalck and Deborshee Sen and David B. Dunson},
  journal = {Journal of the American Statistical Association},
  year = {2026},
  month = jul,
  doi = {10.1080/01621459.2026.2688950},
  url = {https://doi.org/10.1080/01621459.2026.2688950},
  note = {CC BY},
}
RIS
TY  - JOUR
TI  - Scalable Bayesian Inference for Time Series via Divide and Conquer
AU  - Rihui Ou
AU  - Lachlan Astfalck
AU  - Deborshee Sen
AU  - David B. Dunson
PY  - 2026
DA  - 2026/07/21
JO  - Journal of the American Statistical Association
DO  - 10.1080/01621459.2026.2688950
UR  - https://doi.org/10.1080/01621459.2026.2688950
AB  - Bayesian computation often scales poorly with increasing data size, motivating developments such as divide-and-conquer approaches for scalable inference. These methods partition the data into subsets, perform parallel inference on each subset, and aggregate the results into a single posterior. Appealing theoretical properties and practical performance have been demonstrated for independent data; however, methods for dependent data remain challenging. Existing methods rely on ad hoc approximations with limited theoretical guarantees that lead to potentially poor accuracy in practice. Here, we focus on time-series data and propose a simple, scalable divide-and-conquer method for dependent time series, with theoretically rigorous accuracy guarantees. Simulation studies and real-data examples are used to empirically verify the effectiveness of our approach.
SN  - 0162-1459
SN  - 1537-274X
ER  -