{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"institution":[{"name":"medRxiv"}],"indexed":{"date-parts":[[2026,7,25]],"date-time":"2026-07-25T15:03:07Z","timestamp":1784991787378,"version":"3.55.0"},"posted":{"date-parts":[[2026,7,22]]},"group-title":"Infectious Diseases (except HIV\/AIDS)","reference-count":47,"publisher":"openRxiv","license":[{"start":{"date-parts":[[2026,7,22]],"date-time":"2026-07-22T00:00:00Z","timestamp":1784678400000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0\/"}],"funder":[{"DOI":"10.13039\/100000002","name":"National Institutes of Health","doi-asserted-by":"crossref","award":["R35GM156799"],"award-info":[{"award-number":["R35GM156799"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"crossref"}]},{"DOI":"10.13039\/100000002","name":"National Institutes of Health","doi-asserted-by":"crossref","award":["R01AI63023"],"award-info":[{"award-number":["R01AI63023"]}],"id":[{"id":"10.13039\/100000002","id-type":"DOI","asserted-by":"crossref"}]},{"DOI":"10.13039\/100000001","name":"U.S. National Science Foundation","doi-asserted-by":"crossref","award":["DMS-2229605"],"award-info":[{"award-number":["DMS-2229605"]}],"id":[{"id":"10.13039\/100000001","id-type":"DOI","asserted-by":"crossref"}]},{"DOI":"10.13039\/100006474","name":"Columbia University","doi-asserted-by":"crossref","award":["Research Stabilization Fund"],"award-info":[{"award-number":["Research Stabilization Fund"]}],"id":[{"id":"10.13039\/100006474","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"accepted":{"date-parts":[[2026,7,22]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                <jats:p>Real-time forecasts of seasonal respiratory outbreaks are critical for public health preparedness and healthcare planning. Multi-model ensembles, which combine predictions from individual models, have become a leading approach for operational outbreak forecasting. Their success, however, depends in part on the assumption that component models make sufficiently independent errors. Here, we examined this assumption using archived real-time forecasts for influenza hospitalizations and influenza-like illness (ILI) in the United States. We found that component models with diverse structures and calibration methods shared systematic forecast errors during epidemic growth and around epidemic peaks, reflecting the common challenge of tracking rapid changes in epidemic dynamics from real-time surveillance data. Because such shared errors cannot be fully corrected by ensembling alone, we developed a deep learning framework that learns structured residual errors from historical forecasts and uses them to correct ensemble predictions. This framework improved influenza hospitalization forecasts across horizons and geographic scales, reducing the Weighted Interval Score by up to 20% at the national level and 12% across states relative to official ensemble forecasts, with the largest improvements at the near-term horizon and during epidemic growth and peak periods. We further showed that learned residual structures transferred across ensembles formed from different component models, making the approach robust to changes in model participation across seasons. The framework also improved ensemble forecasts for ILI, although gains were more modest. These findings reveal a fundamental challenge in ensemble forecasting and provide a generalizable approach for improving real-time epidemic forecasts.<\/jats:p>","DOI":"10.64898\/2026.07.20.26358508","type":"posted-content","created":{"date-parts":[[2026,7,23]],"date-time":"2026-07-23T01:25:23Z","timestamp":1784769923000},"source":"Crossref","is-referenced-by-count":0,"title":["Learning shared forecast-error structure to improve ensemble forecasts of seasonal respiratory outbreaks"],"prefix":"10.64898","author":[{"ORCID":"https:\/\/orcid.org\/0009-0004-3498-5538","authenticated-orcid":false,"given":"Yuchen","family":"Qin","sequence":"first","affiliation":[{"name":"Department of Systems & Information Engineering, University of Virginia, Charlottesville, VA, USA"},{"name":"Department of Environmental Health Sciences, Mailman School of Public Health, Columbia University,New York, NY,USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7008-2943","authenticated-orcid":false,"given":"Hongru","family":"Du","sequence":"additional","affiliation":[{"name":"Department of Systems & Information Engineering, University of Virginia, Charlottesville, VA, USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7072-2995","authenticated-orcid":false,"given":"Sen","family":"Pei","sequence":"additional","affiliation":[{"name":"Department of Environmental Health Sciences, Mailman School of Public Health, Columbia University,New York, NY,USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"54368","reference":[{"issue":"1","key":"2026072507500800000_2026.07.20.26358508v1.1","doi-asserted-by":"crossref","first-page":"12406","DOI":"10.1038\/s41598-018-30378-w","article-title":"Assessing the use of influenza forecasts and epidemiological modeling in public health decision making in the united states","volume":"8","year":"2018","journal-title":"Scientific reports"},{"issue":"1","key":"2026072507500800000_2026.07.20.26358508v1.2","doi-asserted-by":"crossref","first-page":"6289","DOI":"10.1038\/s41467-024-50601-9","article-title":"Evaluation of flusight influenza forecasting in the 2021\u201322 and 2022\u201323 seasons with a new target laboratory-confirmed influenza hospitalizations","volume":"15","year":"2024","journal-title":"Nature communications"},{"key":"2026072507500800000_2026.07.20.26358508v1.3","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.1812594116"},{"key":"2026072507500800000_2026.07.20.26358508v1.4","doi-asserted-by":"crossref","unstructured":"Nicholas G Reich , Justin Lessler , Sebastian Funk , Cecile Viboud , Alessandro Vespignani , Ryan J Tibshirani , Katriona Shea , Melanie Schienle , Michael C Runge , Roni Rosenfeld , et al. Collaborative hubs: making the most of predictive epidemic modeling, 2022.","DOI":"10.2105\/AJPH.2022.306831"},{"issue":"11","key":"2026072507500800000_2026.07.20.26358508v1.5","doi-asserted-by":"crossref","first-page":"e1007486","DOI":"10.1371\/journal.pcbi.1007486","article-title":"Accuracy of real-time multi-model ensemble forecasts for seasonal influenza in the us","volume":"15","year":"2019","journal-title":"PLoS computational biology"},{"issue":"144","key":"2026072507500800000_2026.07.20.26358508v1.6","doi-asserted-by":"crossref","first-page":"20180174","DOI":"10.1098\/rsif.2018.0174","article-title":"Evaluation of mechanistic and statistical methods in forecasting influenza-like illness","volume":"15","year":"2018","journal-title":"Journal of The Royal Society Interface"},{"key":"2026072507500800000_2026.07.20.26358508v1.7","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.2113561119"},{"key":"2026072507500800000_2026.07.20.26358508v1.8","unstructured":"Centers for Disease Control and Prevention. FluSight 2023\u20132024 evaluation. https:\/\/www.cdc.gov\/flu-forecasting\/evaluation\/2023-2024-report.html, 2025."},{"key":"2026072507500800000_2026.07.20.26358508v1.9","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.1208772109"},{"key":"2026072507500800000_2026.07.20.26358508v1.10","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.2508575122"},{"issue":"6","key":"2026072507500800000_2026.07.20.26358508v1.11","doi-asserted-by":"crossref","first-page":"e1006134","DOI":"10.1371\/journal.pcbi.1006134","article-title":"Nonmechanistic forecasts of seasonal influenza with iterative one-week-ahead distributions","volume":"14","year":"2018","journal-title":"PLoS computational biology"},{"key":"2026072507500800000_2026.07.20.26358508v1.12","doi-asserted-by":"publisher","DOI":"10.1126\/sciadv.abb1237"},{"key":"2026072507500800000_2026.07.20.26358508v1.13","doi-asserted-by":"crossref","unstructured":"Sasikiran Kandula , Sen Pei , and Jeffrey Shaman . Improved forecasts of influenza-associated hospitalization rates with google search trends. Journal of the Royal Society Interface, 16(155), 2019.","DOI":"10.1098\/rsif.2019.0080"},{"key":"2026072507500800000_2026.07.20.26358508v1.14","doi-asserted-by":"publisher","DOI":"10.1038\/nature07634"},{"key":"2026072507500800000_2026.07.20.26358508v1.15","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.1515373112"},{"key":"2026072507500800000_2026.07.20.26358508v1.16","doi-asserted-by":"publisher","DOI":"10.7554\/eLife.104282"},{"key":"2026072507500800000_2026.07.20.26358508v1.17","doi-asserted-by":"crossref","unstructured":"Hongru Du , Ensheng Dong , Hamada S Badr , Mary E Petrone , Nathan D Grubaugh , and Lauren M Gardner . Incorporating variant frequencies data into short-term forecasting for covid-19 cases and deaths in the usa: a deep learning approach. EBioMedicine, 89, 2023.","DOI":"10.1016\/j.ebiom.2023.104482"},{"issue":"10","key":"2026072507500800000_2026.07.20.26358508v1.18","doi-asserted-by":"crossref","first-page":"e1008301","DOI":"10.1371\/journal.pcbi.1008301","article-title":"Aggregating forecasts of multiple respiratory pathogens supports more accurate forecasting of influenza-like illness","volume":"16","year":"2020","journal-title":"PLoS computational biology"},{"key":"2026072507500800000_2026.07.20.26358508v1.19","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.1708856115"},{"issue":"1","key":"2026072507500800000_2026.07.20.26358508v1.20","doi-asserted-by":"crossref","first-page":"726","DOI":"10.1038\/s41467-021-21018-5","article-title":"Forecasting influenza activity using machine-learned mobility map","volume":"12","year":"2021","journal-title":"Nature communications"},{"issue":"1","key":"2026072507500800000_2026.07.20.26358508v1.21","doi-asserted-by":"crossref","first-page":"222","DOI":"10.1038\/s41467-020-20399-3","article-title":"Optimizing respiratory virus surveillance networks using uncertainty propagation","volume":"12","year":"2021","journal-title":"Nature Communications"},{"issue":"5","key":"2026072507500800000_2026.07.20.26358508v1.22","doi-asserted-by":"crossref","first-page":"e340","DOI":"10.1016\/S2666-5247(22)00386-X","article-title":"Wastewater concentrations of human influenza, metapneumovirus, parainfluenza, respiratory syncytial virus, rhinovirus, and seasonal coronavirus nucleic-acids during the covid-19 pandemic: a surveillance study","volume":"4","year":"2023","journal-title":"The Lancet Microbe"},{"issue":"3","key":"2026072507500800000_2026.07.20.26358508v1.23","doi-asserted-by":"crossref","first-page":"1366","DOI":"10.1016\/j.ijforecast.2022.06.005","article-title":"Comparing trained and untrained probabilistic ensemble forecasts of covid-19 cases and deaths in the united states","volume":"39","year":"2023","journal-title":"International journal of forecasting"},{"issue":"2","key":"2026072507500800000_2026.07.20.26358508v1.24","doi-asserted-by":"crossref","first-page":"e1005910","DOI":"10.1371\/journal.pcbi.1005910","article-title":"Prediction of infectious disease epidemics via weighted density ensembles","volume":"14","year":"2018","journal-title":"PLoS computational biology"},{"issue":"2","key":"2026072507500800000_2026.07.20.26358508v1.25","doi-asserted-by":"crossref","first-page":"675","DOI":"10.1007\/s00477-024-02889-8","article-title":"Epiforecaster: a novel deep learning ensemble optimization approach to combining forecasts for emerging epidemic out-breaks","volume":"39","year":"2025","journal-title":"Stochastic Environmental Research and Risk Assessment"},{"key":"2026072507500800000_2026.07.20.26358508v1.26","doi-asserted-by":"publisher","DOI":"10.1016\/j.epidem.2019.100379"},{"key":"2026072507500800000_2026.07.20.26358508v1.27","doi-asserted-by":"publisher","DOI":"10.1002\/sim.9219"},{"key":"2026072507500800000_2026.07.20.26358508v1.28","doi-asserted-by":"crossref","unstructured":"Shaochong Xu , Hongru Du , Ensheng Dong , Xianglong Wang , Liyue Zhang , and Lauren M Gardner . An optimized geo-hierarchical ensemble model to forecast hospitalizations from respiratory viruses in the united states. Epidemics, page 100869, 2025.","DOI":"10.1016\/j.epidem.2025.100869"},{"issue":"1","key":"2026072507500800000_2026.07.20.26358508v1.29","doi-asserted-by":"crossref","first-page":"8625","DOI":"10.1038\/s41467-024-52504-1","article-title":"An adaptive weight ensemble approach to forecast influenza activity in an irregular seasonality context","volume":"15","year":"2024","journal-title":"Nature Communications"},{"key":"2026072507500800000_2026.07.20.26358508v1.30","doi-asserted-by":"publisher","DOI":"10.64898\/2026.06.26.26356648"},{"key":"2026072507500800000_2026.07.20.26358508v1.31","doi-asserted-by":"publisher","DOI":"10.1002\/sim.9884"},{"key":"2026072507500800000_2026.07.20.26358508v1.32","doi-asserted-by":"crossref","first-page":"15647","DOI":"10.1609\/aaai.v37i13.26855","article-title":"Phase-informed bayesian ensemble models improve performance of covid-19 forecasts","volume":"37","year":"2023","journal-title":"Proceedings of the AAAI Conference on Artificial Intelligence"},{"issue":"9","key":"2026072507500800000_2026.07.20.26358508v1.33","first-page":"1967","article-title":"Optimizing disease outbreak forecast ensembles","volume":"30","year":"2024","journal-title":"Emerging infectious diseases"},{"key":"2026072507500800000_2026.07.20.26358508v1.34","article-title":"Not all accuracy is equal: Prioritizing independence in infectious disease forecasting","year":"2025","journal-title":"arXiv preprint"},{"issue":"4","key":"2026072507500800000_2026.07.20.26358508v1.35","doi-asserted-by":"crossref","first-page":"e13403","DOI":"10.2196\/13403","article-title":"Differences in regional patterns of influenza activity across surveillance systems in the united states: comparative evaluation","volume":"5","year":"2019","journal-title":"JMIR public health and surveillance"},{"issue":"3","key":"2026072507500800000_2026.07.20.26358508v1.36","doi-asserted-by":"crossref","first-page":"e1006020","DOI":"10.1371\/journal.pcbi.1006020","article-title":"Deploying digital health data to optimize influenza surveillance at national and local scales","volume":"14","year":"2018","journal-title":"PLoS computational biology"},{"key":"2026072507500800000_2026.07.20.26358508v1.37","doi-asserted-by":"publisher","DOI":"10.1093\/biomet\/53.3-4.325"},{"key":"2026072507500800000_2026.07.20.26358508v1.38","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1997.9.8.1735"},{"key":"2026072507500800000_2026.07.20.26358508v1.39","article-title":"Graph attention networks","year":"2017","journal-title":"arXiv preprint"},{"issue":"11","key":"2026072507500800000_2026.07.20.26358508v1.40","doi-asserted-by":"crossref","first-page":"e1013709","DOI":"10.1371\/journal.pcbi.1013709","article-title":"Real-time forecasting of data revisions in epidemic surveillance streams","volume":"21","year":"2025","journal-title":"PLOS Computational Biology"},{"key":"2026072507500800000_2026.07.20.26358508v1.41","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-96709-3"},{"key":"2026072507500800000_2026.07.20.26358508v1.42","doi-asserted-by":"crossref","unstructured":"Kaiming He , Xiangyu Zhang , Shaoqing Ren , and Jian Sun . Deep residual learning for image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 770\u2013778, 2016.","DOI":"10.1109\/CVPR.2016.90"},{"key":"2026072507500800000_2026.07.20.26358508v1.43","doi-asserted-by":"publisher","DOI":"10.1214\/aos\/1013203451"},{"key":"2026072507500800000_2026.07.20.26358508v1.44","doi-asserted-by":"publisher","DOI":"10.1214\/18-BA1117"},{"issue":"2","key":"2026072507500800000_2026.07.20.26358508v1.45","doi-asserted-by":"crossref","first-page":"e2518051123","DOI":"10.1073\/pnas.2518051123","article-title":"Reconstructing the early spatial spread of pandemic respiratory viruses in the united states","volume":"123","year":"2026","journal-title":"Proceedings of the National Academy of Sciences"},{"issue":"2","key":"2026072507500800000_2026.07.20.26358508v1.46","doi-asserted-by":"crossref","first-page":"e1008618","DOI":"10.1371\/journal.pcbi.1008618","article-title":"Evaluating epidemic forecasts in an interval format","volume":"17","year":"2021","journal-title":"PLoS computational biology"},{"key":"2026072507500800000_2026.07.20.26358508v1.47","unstructured":"Carnegie Mellon University Delphi Group. Delphi Epidata API: NHSN Hospital Respiratory Data. https:\/\/cmu-delphi.github.io\/delphi-epidata\/, 2025. Accessed 2025."}],"container-title":[],"original-title":[],"link":[{"URL":"https:\/\/syndication.highwire.org\/content\/doi\/10.64898\/2026.07.20.26358508","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,25]],"date-time":"2026-07-25T14:50:20Z","timestamp":1784991020000},"score":1,"resource":{"primary":{"URL":"http:\/\/medrxiv.org\/lookup\/doi\/10.64898\/2026.07.20.26358508"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,7,22]]},"references-count":47,"URL":"https:\/\/doi.org\/10.64898\/2026.07.20.26358508","relation":{},"subject":[],"published":{"date-parts":[[2026,7,22]]},"subtype":"preprint"}}