A disease-agnostic approach to ensemble learning for infectious disease forecasting

2026-03-20 · Source: Machine learning : nature.com subject feeds · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics, Epidemiological Forecasting · Depth: Advanced, long

Summary

A new ensembling strategy, epiFFORMA, has been developed for infectious disease forecasting that operates without relying on extensive historical data for specific diseases, making it disease-agnostic. This approach addresses a critical challenge in public health by enabling accurate real-time forecasting for emerging diseases where historical data is scarce. The epiFFORMA model builds upon the FFORMA model, originally from the M4 forecasting competition, by incorporating epidemiological dynamics through synthetic data generation. Researchers demonstrated that epiFFORMA outperforms a naive, equal-weighting ensembling strategy and individual component models across outbreaks of COVID-19, diphtheria, influenza-like illness, dengue, measles, mumps, polio, rubella, smallpox, and chikungunya. All data used in this study, including COVID-19 case data from Johns Hopkins University and ILI data from the U.S. CDC, were publicly available.

Key takeaway

For research scientists developing public health interventions, epiFFORMA offers a robust method to forecast emerging infectious diseases, even with limited historical data. You should consider integrating this disease-agnostic ensembling strategy into your predictive models to enhance real-time outbreak response and reduce morbidity and mortality. This approach provides a significant advantage over traditional methods that require extensive historical data, which is often unavailable for novel pathogens.

Key insights

EpiFFORMA enables accurate, disease-agnostic infectious disease forecasting by using synthetic data to train ensemble models.

Principles

• Synthetic data can overcome historical data scarcity.
• Ensemble models generally outperform individual components.
• Disease-agnostic methods enhance preparedness for novel pathogens.

Method

The epiFFORMA model determines component weights for an ensemble without historical data by building on the FFORMA model and integrating epidemiological dynamics via synthetic data. This allows for robust forecasting across diverse infectious diseases.

In practice

• Apply epiFFORMA for emerging disease outbreak prediction.
• Utilize synthetic data to augment sparse real-world datasets.
• Combine multiple models for improved forecast accuracy.

Topics

• Infectious Disease Forecasting
• Ensemble Learning
• Disease-Agnostic Models
• Synthetic Data
• FFORMA Model

Code references

• CSSEGISandData/COVID-19_Unified-Dataset
• wmarciel/Chikungunya-in-Brazil-2013-2022

Best for: Research Scientist, AI Researcher, AI Scientist, Data Scientist

Related on AIssential

• Hybrid Probabilistic Forecasting of Under-Five Malaria Admissions in Ghana: A Gaussian Process Regression with Holt-Winters Smoothing — Combining GPR with Holt-Winters provides robust probabilistic malaria forecasting, capturing non-linearity and seasonality.
• Prospective multi-pathogen disease forecasting using autonomous LLM-guided tree search — LLM-guided tree search autonomously generates and optimizes disease forecasting models, matching human expert performance.
• EpiEvolve: Self-Evolving Agents for Streaming Pandemic Forecasting under Regime Shifts — LLM forecasters can adapt to streaming data and regime shifts via memory-based self-evolution, not parameter updates.
• Doubly Outlier-Robust Online Infinite Hidden Markov Model — BR-iHMM offers robust online learning for streaming data with outliers and model misspecification.
• Benchmarking Counterfactual Prediction in Epidemic Time Series with Time-Varying Interventions — A new benchmark provides realistic, observable counterfactuals for epidemic time series under dynamic interventions, addressing a critical gap.
• Embedding-Based Federated Learning with Runtime Governance for Iron Deficiency Prediction — Personalized aggregation and runtime governance are crucial for effective, deployable healthcare federated learning with heterogeneous data.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Machine learning : nature.com subject feeds.