A1015
Title: Fast and efficient inference for flexible spatial extremes models
Authors: Boris Beranger - University of New South Wales (Australia) [presenting]
Scott Sisson - University of New South Wales (Austria)
Peng Zhong - University of New South Wales (Australia)
Abstract: Statistical modelling of spatial extreme events has gained increasing attention over the last few decades, with max-stable processes and, more recently, r-Pareto processes becoming the reference tools for the statistical analysis of asymptotically dependent data. Although inference for r-Pareto processes is easier than for max-stable processes, there remain major hurdles for their application to very high-dimensional datasets within a reasonable timeframe. In addition, both approaches have almost exclusively considered the Brown-Resnick model for its Gaussian-based foundations and the continuity of its exponent measure. A class of models is derived, for which this continuity property holds, and the skewed Brown-Resnick model is presented, an extension of the Brown-Resnick that allows for non-stationarity in the dependence structure, and the truncated extremal-t, a refinement of the well-known extremal-t model. An inference methodology is used based on the intensity function of the process, which is derived from the exponent measure, and the statistical and computational efficiency of this approach is demonstrated. Applications to two real-world problems illustrate valuable gains in flexibility from the proposed models as well as appealing computational gains over reference methodologies.
