Statistics seminar - Distinguished Professor Noel Cressie - University of Wollongong

A seminar by Distinguished Professor Noel Cressie from University of Wollongong

Title: Spatial prediction of non-negative spatial processes using asymmetric losses

Abstract: A major component of inference in spatial statistics is that of spatial prediction of an unknown value of a latent spatial process, based on noisy measurements of the process taken at various locations in a spatial domain. The most commonly used predictor is the conditional expectation of the unknown value given the data. By considering the spatial-prediction problem from a decision-theoretic viewpoint, one can recognise the conditional expectation as optimal for squared-error loss (SEL). In this talk, we consider spatial prediction of processes that take non-negative values. A family of power-divergence loss functions, which is indexed by the choice of a power parameter, is well defined for predictor and predictand when the spatial process is non-negative, and the power parameter controls the asymmetry of the losses. Taking a hierarchical spatial-statistical-modelling approach, it can be seen that this class of asymmetric loss functions generates new optimal spatial predictors. An application is given to spatial prediction of zinc concentrations in soil on a floodplain of the Meuse River in the Netherlands. This talk is based on joint research with Alan Pearse and David Gunawan at the University of Wollongong.

For further information, please contact RSFAS Seminars.

All information collected by the University is governed by the ANU Privacy Policy.

Start Date
End Date
PAP Moran G058
Distinguished Professor Noel Cressie