The MELBS team winning entry for the EVA2017 competition for spatiotemporal prediction of extreme rainfall using generalized extreme value quantiles
- 121 Downloads
We present our winning entry for the EVA2017 challenge on spatiotemporal prediction of extreme precipitation. The aim of the competition is to predict extreme rainfall quantiles that score as low as possible on the competition error metric. Good or bad predictions are defined only by the metric used. Our methodology was simple and produced accurate predictions under this metric. This outcome emphasizes the importance of cross-validation and identifying model over-fitting.
KeywordsData mining Extreme rainfall Generalized extreme value distribution Spatiotemporal prediction
Unable to display preview. Download preview PDF.
We thank the organizing committee of the 10th international conference on Extreme Value Analysis, and Olivier Wintenberger for organizing the prediction challenge. Laleh Tafakori and Kate Saunders would like to thank the Australian Research Council for supporting this work through Laureate Fellowship FL130100039. The authors also acknowledge the support of The Australian Research Council Center of Excellence for Mathematical and Statistical Frontiers.
- Hyndman, R.J., Athanasopoulos, G.: Forecasting: Principles and Practice. OTexts (2013)Google Scholar
- Wintenberger, O.: Editorial: special issue on the Extreme Value Analysis conference challenge “Prediction of extremal precipitation”. Extremes. To Appear (2018)Google Scholar
- Witten, I.H., Frank, E., Hall, M.A., Pal, C.J.: Data Mining: Practical Machine Learning Tools and Techniques. Morgan Kaufmann, Cambridge (2017)Google Scholar