|
Abstract:
|
A common problem in spatial statistics is to predict exceedances and the regions of the exceedances. This has application in environmental sciences, natural resources, and agriculture, since rare events tend to have a strong impact on environment. We develop a method involving a loss function based on Baddeley's (1992) metric for binary images. The predictor based on the loss function is obtained by minimizing the posterior expected loss when the spatial trend, noise, and spatial-covariance parameters are estimated; in practice, minimization is achieved by simulated annealing. Through simulation, we calibrate the choice of a tuning parameter in the loss function, and investigate the predictor's performance. We then apply our methodology to a spatial dataset of decadal temperature changes over the Americas.
|
