{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,21]],"date-time":"2026-08-21T18:30:30Z","timestamp":1787337030740,"version":"build-2736575974"},"reference-count":31,"publisher":"Public Library of Science (PLoS)","issue":"8","license":[{"start":{"date-parts":[[2026,8,21]],"date-time":"2026-08-21T00:00:00Z","timestamp":1787270400000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["www.plosone.org"],"crossmark-restriction":false},"short-container-title":["PLoS One"],"abstract":"<jats:p>Non-stationary time series are common in many real-world domains, including infectious disease spread, where the underlying relationships between variables evolve over time. However, most existing forecasting methods assume stationarity and fail to capture changing causal dynamics. To address this challenge, we propose the Causal Regime Bayesian (CaReBayes) forecasting framework, which integrates regime detection, causal discovery, and Bayesian forecasting within a unified approach. CaReBayes segments time series into regimes using temporal causal discovery, fits a Bayesian structural autoregressive model for each regime, classifies the current regime, then performs regime-specific forecasting with uncertainty quantification. The framework introduces methodological advances: a grid-search procedure for automated regime-dependent causal discovery, a classification method that assigns future observations to regimes based on learned Bayesian structures, and regime-conditioned Bayesian structural forecasting. Across both synthetic and Ontario COVID-19 time series data, CaReBayes outperforms benchmark models for time series forecasting. In addition to improved forecasting performance, it produces regime-dependent causal graphs that summarize candidate structural relationships in the system, enhancing interpretability.<\/jats:p>","DOI":"10.1371\/journal.pone.0356474","type":"journal-article","created":{"date-parts":[[2026,8,21]],"date-time":"2026-08-21T17:41:26Z","timestamp":1787334086000},"page":"e0356474","update-policy":"https:\/\/doi.org\/10.1371\/journal.pone.corrections_policy","source":"Crossref","is-referenced-by-count":0,"title":["Regime-aware causal Bayesian forecasting for non-stationary time series"],"prefix":"10.1371","volume":"21","author":[{"ORCID":"https:\/\/orcid.org\/0009-0000-6017-8079","authenticated-orcid":true,"given":"Brandon","family":"Mossop","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Salimur","family":"Choudhury","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"340","published-online":{"date-parts":[[2026,8,21]]},"reference":[{"issue":"4","key":"pone.0356474.ref001","first-page":"675","article-title":"Causality: models, reasoning, and inference","volume":"19","author":"LG Neuberg","year":"2003","journal-title":"Econometric Theory"},{"issue":"3","key":"pone.0356474.ref002","doi-asserted-by":"crossref","first-page":"705","DOI":"10.1016\/j.ijforecast.2021.11.001","article-title":"Forecasting: theory and practice","volume":"38","author":"F Petropoulos","year":"2022","journal-title":"International Journal of Forecasting"},{"issue":"2","key":"pone.0356474.ref003","doi-asserted-by":"crossref","first-page":"357","DOI":"10.2307\/1912559","article-title":"A New Approach to the Economic Analysis of Nonstationary Time Series and the Business Cycle","volume":"57","author":"JD Hamilton","year":"1989","journal-title":"Econometrica"},{"key":"pone.0356474.ref004","doi-asserted-by":"crossref","first-page":"6","DOI":"10.1007\/978-3-642-51684-9_2","article-title":"The markov-switching vector autoregressive model","volume-title":"Markov-switching vector autoregressions: Modelling, statistical inference, and application to business cycle analysis","author":"HM Krolzig","year":"1997"},{"key":"pone.0356474.ref005","doi-asserted-by":"crossref","DOI":"10.1017\/CBO9781107049994","volume-title":"Forecasting, structural time series models and the Kalman filter","author":"AC Harvey","year":"1990"},{"issue":"1","key":"pone.0356474.ref006","doi-asserted-by":"crossref","first-page":"49","DOI":"10.1002\/for.859","article-title":"Subset threshold autoregression","volume":"22","author":"MKP So","year":"2003","journal-title":"Journal of Forecasting"},{"issue":"2","key":"pone.0356474.ref007","doi-asserted-by":"crossref","first-page":"191","DOI":"10.1080\/00949655.2018.1540619","article-title":"On hysteretic vector autoregressive model with applications","volume":"89","author":"CWS Chen","year":"2018","journal-title":"Journal of Statistical Computation and Simulation"},{"issue":"4","key":"pone.0356474.ref008","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1145\/2523813","article-title":"A survey on concept drift adaptation","volume":"46","author":"J Gama","year":"2014","journal-title":"ACM Comput Surv"},{"key":"pone.0356474.ref009","unstructured":"Adams RP, MacKay DJ. Bayesian online changepoint detection. In: 2007. https:\/\/arxiv.org\/abs\/0710.3742"},{"issue":"4","key":"pone.0356474.ref010","doi-asserted-by":"crossref","first-page":"1569","DOI":"10.1109\/TSP.2010.2102756","article-title":"Bayesian Nonparametric Inference of Switching Dynamic Linear Models","volume":"59","author":"E Fox","year":"2011","journal-title":"IEEE Trans Signal Process"},{"key":"pone.0356474.ref011","doi-asserted-by":"crossref","unstructured":"Zhang K, Huang B, Zhang J, Glymour C, Sch\u00f6lkopf B. Causal discovery from nonstationary\/heterogeneous data: Skeleton estimation and orientation determination. In: IJCAI: Proceedings of the Conference, 2017. 1347.","DOI":"10.24963\/ijcai.2017\/187"},{"issue":"11","key":"pone.0356474.ref012","doi-asserted-by":"crossref","DOI":"10.1063\/5.0020538","article-title":"Reconstructing regime-dependent causal relationships from observational time series","volume":"30","author":"E Saggioro","year":"2020","journal-title":"Chaos: An Interdisciplinary Journal of Nonlinear Science"},{"key":"pone.0356474.ref013","unstructured":"Huang B, Zhang K, Gong M, Glymour C. Causal discovery and forecasting in nonstationary environments with state-space models. In: International conference on machine learning, 2019. 2901\u201310."},{"key":"pone.0356474.ref014","doi-asserted-by":"crossref","first-page":"107827","DOI":"10.1016\/j.csda.2023.107827","article-title":"Bayesian modeling of spatial integer-valued time series","volume":"188","author":"CWS Chen","year":"2023","journal-title":"Computational Statistics & Data Analysis"},{"issue":"6","key":"pone.0356474.ref015","doi-asserted-by":"crossref","first-page":"2119","DOI":"10.1007\/s00477-024-02671-w","article-title":"Spatial-temporal hurdle model vs. spatial zero-inflated GARCH model: analysis of weekly dengue fever cases","volume":"38","author":"CWS Chen","year":"2024","journal-title":"Stoch Environ Res Risk Assess"},{"issue":"11","key":"pone.0356474.ref016","doi-asserted-by":"crossref","DOI":"10.1126\/sciadv.aau4996","article-title":"Detecting and quantifying causal associations in large nonlinear time series datasets","volume":"5","author":"J Runge","year":"2019","journal-title":"Sci Adv"},{"issue":"1","key":"pone.0356474.ref017","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1021\/ci0342472","article-title":"The problem of overfitting","volume":"44","author":"DM Hawkins","year":"2004","journal-title":"J Chem Inf Comput Sci"},{"key":"pone.0356474.ref018","volume-title":"Causation, prediction, and search","author":"P Spirtes","year":"2000"},{"key":"pone.0356474.ref019","doi-asserted-by":"crossref","first-page":"2343","DOI":"10.1080\/02664763.2023.2177625","article-title":"On regime changes of COVID-19 outbreak","volume":"50","author":"A Tchorbadjieff","year":"2023","journal-title":"J Appl Stat"},{"key":"pone.0356474.ref020","article-title":"Regime shifts in the COVID-19 case fatality rate dynamics: A Markov-switching autoregressive model analysis. Chaos, Solitons & Fractals:","volume":"6","author":"YA Shiferaw","year":"2021","journal-title":"X"},{"key":"pone.0356474.ref021","doi-asserted-by":"crossref","unstructured":"Mossop B, Rahman QA. Infectious Disease Forecasting using Multivariate Incomplete Time-series: A Hybrid Architecture with Stacked Dilated Causal Convolutions. In: 2023 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), 2023. 4220\u20137. https:\/\/doi.org\/10.1109\/bibm58861.2023.10385253","DOI":"10.1109\/BIBM58861.2023.10385253"},{"key":"pone.0356474.ref022","unstructured":"Health O. Confirmed positive cases of COVID-19 in Ontario. 2021 [cited 2025 October 6]. https:\/\/data.ontario.ca\/dataset\/confirmed-positive-cases-of-covid-19-in-ontario"},{"issue":"12","key":"pone.0356474.ref023","doi-asserted-by":"crossref","DOI":"10.1371\/journal.pcbi.1008409","article-title":"Practical considerations for measuring the effective reproductive number, Rt","volume":"16","author":"KM Gostic","year":"2020","journal-title":"PLoS Comput Biol"},{"key":"pone.0356474.ref024","unstructured":"Health O. Effective reproduction number (Re) for COVID-19 in Ontario. 2021 [cited 2025 October 6]. https:\/\/data.ontario.ca\/dataset\/effective-reproduction-number-re-for-covid-19-in-ontario"},{"key":"pone.0356474.ref025","unstructured":"LLC G. Google COVID-19 Community Mobility Reports. 2022 [cited 2025 October 6]. https:\/\/www.google.com\/covid19\/mobility"},{"issue":"1","key":"pone.0356474.ref026","doi-asserted-by":"crossref","first-page":"92","DOI":"10.25080\/Majora-92bf1922-011","article-title":"Statsmodels: econometric and statistical modeling with python","volume":"7","author":"S Seabold","year":"2010","journal-title":"SciPy"},{"key":"pone.0356474.ref027","doi-asserted-by":"crossref","first-page":"245","DOI":"10.1613\/jair.1.12228","article-title":"A Survey on the Explainability of Supervised Machine Learning","volume":"70","author":"N Burkart","year":"2021","journal-title":"jair"},{"issue":"1","key":"pone.0356474.ref028","doi-asserted-by":"crossref","first-page":"18","DOI":"10.3390\/e23010018","article-title":"Explainable AI: A Review of Machine Learning Interpretability Methods","volume":"23","author":"P Linardatos","year":"2020","journal-title":"Entropy (Basel)"},{"key":"pone.0356474.ref029","unstructured":"Doshi-Velez F, Kim B. Towards a rigorous science of interpretable machine learning. arXiv preprint. 2017. doi: arXiv:1702.08608"},{"issue":"6","key":"pone.0356474.ref030","doi-asserted-by":"crossref","first-page":"442","DOI":"10.1016\/j.tig.2020.03.005","article-title":"Opening the Black Box: Interpretable Machine Learning for Geneticists","volume":"36","author":"CB Azodi","year":"2020","journal-title":"Trends Genet"},{"key":"pone.0356474.ref031","doi-asserted-by":"crossref","unstructured":"Gong C, Yao D, Zhang C, Li W, Bi J, Du L, et al. Causal Discovery from Temporal Data. In: Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2023. 5803\u20134. https:\/\/doi.org\/10.1145\/3580305.3599552","DOI":"10.1145\/3580305.3599552"}],"container-title":["PLOS One"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0356474","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,21]],"date-time":"2026-08-21T17:41:32Z","timestamp":1787334092000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0356474"}},"subtitle":[],"editor":[{"given":"Cathy W. S.","family":"Chen","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2026,8,21]]},"references-count":31,"journal-issue":{"issue":"8","published-online":{"date-parts":[[2026,8,21]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pone.0356474","relation":{},"ISSN":["1932-6203"],"issn-type":[{"value":"1932-6203","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,8,21]]}}}