The abstract is read at the publisher; the summary is JClub's.
GRU, Transformer, and Chronos models forecast weekly influenza A positivity-rate incidence across eight Canadian provinces with lower MAE than the CEM in four provinces during 2025/2026.
- Why it matters: Reliable seasonal epidemic forecasts are difficult, and it is unclear whether AI approaches improve accuracy enough to justify their complexity compared with established statistical and compartmental models.
- What they did: The study independently generated up-to-5-week forecasts for each province from October 11, 2025 to March 28, 2026, using weekly positivity-rate incidence of seasonal influenza A, and evaluated MAE.
- The result: CEM and ARIMA beat all AI models in three and one provinces, respectively; Prophet had the highest MAE, PINN was worse than CEM, and AI was comparable or slightly better overall.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.