Understanding Key Features of Time Series Foundation Models from Epidemic Forecasting

2026-06-17

AI Summary

This study addresses the lack of systematic evaluation of modern time series models for influenza forecasting, which hinders precise public health decision-making. The authors conduct a comprehensive comparison of classical neural networks, numerical Transformers, pretrained time series foundation models, and large language models for regional influenza prediction over horizons of one to four weeks, incorporating hospitalization data as an auxiliary signal. They propose a mixture-of-experts architecture that integrates multiple pretrained predictors, significantly outperforming baseline methods in multi-step forecasting. Their analysis reveals that aligning pretraining objectives with epidemiological dynamics is critical for long-horizon prediction accuracy, that hospitalization data enhances robustness in specific scenarios, and that large language models exhibit limited effectiveness for this task.

Abstract

Seasonal influenza infects millions of people and causes substantial morbidity and mortality in the United States each year, making accurate short-term forecasting a core public-health need. Reliable forecasts of epidemic time series can inform vaccination timing, hospital staffing, and resource allocation, yet the comparative behavior of modern forecasting architectures on infectious-disease surveillance data remains insufficiently characterized. We address this gap through a systematic evaluation of regional influenza forecasting using influenza-like illness surveillance and influenza-associated hospitalization time series under both temporal and spatial generalization settings for 1-4-week-ahead prediction. We compare classical neural network architectures, numerical transformer-based models, pretrained time series foundation models, and LLM-based forecasting approaches. Across tasks, we demonstrate that a mixture-of-experts model that fuses multiple pretrained forecasters achieves the strongest overall performance, indicating that heterogeneous pretrained representations provide complementary predictive information. Our results further show that numerical transformer-based models produce reliable forecasts, while pretraining provides the largest gains at longer horizons, particularly when the pretraining domain is mechanistically aligned with influenza dynamics. In contrast, LLM-based time series methods underperform relative to numerical forecasters in this setting. Finally, we examine hospitalization information as both an auxiliary covariate and a pretraining source. Hospitalization signals provide complementary improvements in selected settings and clarify when additional surveillance streams enhance the robustness of multi-horizon forecasting. These findings provide actionable guidance on model selection, pretraining strategy, and auxiliary-signal use for influenza preparedness.

Problem

• influenza forecasting
• time series foundation models
• epidemic surveillance
• forecasting architectures
• public health

Innovation

• time series foundation models
• mixture-of-experts
• epidemic forecasting
• pretraining alignment
• multi-horizon prediction