B0705
Title: Deep compositional models for nonstationary extremal dependence
Authors: Xuanjie Shao - King Abdullah University of Science and Technology (Saudi Arabia) [presenting]
Jordan Richards - Lancaster University (United Kingdom)
Raphael Huser - King Abdullah University of Science and Technology (Saudi Arabia)
Abstract: Modelling the nonstationarity and anisotropy that often prevail in the extremal dependence of spatial data can be challenging. Inference for stationary and isotropic models is considerably easier, but the assumptions that underpin these models are not typically met by data observed over large or topographically-complex domains. A simple approach to accommodating spatial non-stationarity in Gaussian processes, proposed by a prior study, is to warp the original spatial domain to a latent space where stationarity and isotropy can be reasonably assumed. However, estimating the warping function can be computationally expensive, and the transformation is not guaranteed to be injective, which can lead to physically-unrealistic transformations. A previous study overcame these issues by exploiting deep Gaussian processes, where the transformation is constructed using a deep composition of injective mappings. An extension of this methodology is presented to model non-stationarity in extremal dependence of data by leveraging popularly-applied parametric models for spatial extremal processes.