Author: Michal Ben-Nun; Pete Riley; James Turtle; David P. Bacon; Steven Riley
Title: National and Regional Influenza-Like-Illness Forecasts for the USA Document date: 2018_4_27
ID: cheiabv0_48
Snippet: However, during the middle phase of the season, our models produced less biased 496 estimates and consistently outperformed non-mechanistic models based on the average 497 of historical data. A robust testing of model variants using historical data suggests that 498 spatially coupled models are systematically better than historical NULL models during 499 the middle of the season and are not significantly worse even at the start of the season. 500.....
Document: However, during the middle phase of the season, our models produced less biased 496 estimates and consistently outperformed non-mechanistic models based on the average 497 of historical data. A robust testing of model variants using historical data suggests that 498 spatially coupled models are systematically better than historical NULL models during 499 the middle of the season and are not significantly worse even at the start of the season. 500 We evaluated a simple ensemble and showed that the subjective model choice was 501 better. However, the ranking of individual models suggests that an ensemble of coupled 502 models may outperform our subjective choice. We are considering exactly this 503 experiment for the upcoming season.
Search related documents:
Co phrase search for related documents- individual model and model variant: 1, 2, 3, 4, 5, 6
- individual model and NULL model: 1, 2
- individual model and robust testing: 1
- individual model and simple ensemble: 1, 2, 3
- individual model and subjective choice: 1
- model variant and NULL model: 1
- model variant and subjective choice: 1
- NULL model and simple ensemble: 1
Co phrase search for related documents, hyperlinks ordered by date