Challenge: “Prediction of Extremal Precipitation”Organizer: Olivier Wintenberger, UPMC and Copenhagen University.
We propose a challenge to predict spatio-temporal extremes. The aim of the challenge is to estimate high quantiles and to extrapolate them both in time and space.
The challenge is open to any team by free registration. The best teams will be invited to contribute at a discussion session at EVA 2017 in Delft.
They will also be invited to explain their predictive algorithms in a note that will be published in the journal Extremes
For more details, click here
The ranking is based on the percentage of improvement of the benchmark score on 25% of the test set.
Good luck for producing better final predictions by May 30, 2017!
Name, Challenge 1, Challenge 2
- SNUmultiscale, 68%, 64%
- HuserOpitzRue, 58%, 55%
- Chapell Hill, 51%, 477%
- STOR-i Lancaster, 47%, 43%
- LUCY, 40%, 43%
- SUTD, 34%, 31%
- ISUP, 20%, 29%
- Benchmark, 0%, 0%