2017/12/05 by Felipe Tagle, Stefano Castruccio, Tagle, Felipe +3
Decision Sciences · Economics, Econometrics and Finance · Environmental Science · #Economic and Environmental Valuation #Efficiency Analysis Using DEA #FOS: Computer and information sciences #Methodology (stat.ME) #Soil Geostatistics and Mapping #Spatial and Panel Data Analysis
paper · pdf · doi:10.48550/arxiv.1712.01992
openalex publication_date 2017/12/05 · openalex created_date 2022/09/12 · openalex updated_date 2026/07/28
Large, non-Gaussian spatial datasets pose a considerable modeling challenge\nas the dependence structure implied by the model needs to be captured at\ndifferent scales, while retaining feasible inference. Skew-normal and skew-t\ndistributions have only recently begun to appear in the spatial statistics\nliterature, without much consideration, however, for the ability to capture\ndependence at multiple resolutions, and simultaneously achieve feasible\ninference for increasingly large data sets. This article presents the first\nmulti-resolution spatial model inspired by the skew-t distribution, where a\nlarge-scale effect follows a multivariate normal distribution and the\nfine-scale effects follow a multivariate skew-normal distributions. The\nresulting marginal distribution for each region is skew-t, thereby allowing for\ngreater flexibility in capturing skewness and heavy tails characterizing many\nenvironmental datasets. Likelihood-based inference is performed using a Monte\nCarlo EM algorithm. The model is applied as a stochastic generator of daily\nwind speeds over Saudi Arabia.\n