When and where: Spatiotemporal machine learning forecasts of respiratory infection-related primary care visits

Binay Adhikari

doi:10.1016/j.epidem.2026.100933

When and where: Spatiotemporal machine learning forecasts of respiratory infection-related primary care visits

2026-07-02

0

OA

AI

DOI：10.1016/j.epidem.2026.100933

Original

Original request for help

Save

Abstract

En 中文

• COVID-19 has reshaped respiratory viral dynamics (RSV, influenza, SARS-CoV-2), making seasonal predictions harder for adults, children, and regions. • Syndromic surveillance for viral respiratory infections (VRI) offers timely data on infection trends, but we need more insight into its use for regional forecasting. • We propose spatial and time-lagged features to capture spatiotemporal syndromic patterns, testing them across ML models. • We tested spatial-feature methods (e.g., inverse-distance scaling) and found forecast error varies by syndrome, population, and region size. .

Keywords:

Viral respiratory infection

Forecasting

Syndromic surveillance

Geospatial data

Machine learning

When and where: Spatiotemporal machine learning forecasts of respiratory infection-related primary care visits

Abstract

Journal

Organization