Regression modelling of spatiotemporal extreme U.S. wildfires via partially-interpretable neural networks

arXiv:2208.07581v4 Announce Type: replace-cross
Abstract: Risk management in many environmental settings requires an understanding of the mechanisms that drive extreme events. Useful metrics for quantifying such risk are extreme quantiles of response variables conditioned on predictor variables that describe, e.g., climate, biosphere and environmental states. Typically these quantiles lie outside the range of observable data and so, for estimation, require specification of parametric extreme value models within a regression framework. Classical approaches in this context utilise linear or additive relationships …

abstract arxiv climate cs.lg drive environmental events management metrics modelling networks neural networks regression risk stat.me stat.ml type understanding variables via wildfires