notesum.ai
Published at November 25Machine Learning for the Digital Typhoon Dataset: Extensions to Multiple Basins and New Developments in Representations and Tasks
cs.CV
Released Date: November 25, 2024
Authors: Asanobu Kitamoto1, Erwan Dzik2, Gaspar Faure3
Aff.: 1National Institute of Informatics, Japan; Typhoon Science and Technology Research Center, Yokohama National University, Japan; 2Institut National Polytechnique de Grenoble, France; 3Polytechnique Montréal, Canada

| Forecast Time | t+12 | t+18 | t+24 |
|---|---|---|---|
| Baseline (Persistence forecast) | 9.788 | 13.312 | 16.249 |
| Baseline (TIFS) | ±9.20 | ±10.80 | ±12.00 |
| Time+Position+Pressure | 5.696 | 7.998 | 10.043 |
| Images only | 8.576 | 9.073 | 9.982 |
| Images + Pressure | 5.073 | 6.908 | 8.658 |
| Images + Time + Position + Pressure | 5.240 | 7.006 | 8.671 |