ABSTRACT
The topic of forecasting is discussed in this chapter. Several easy-to-use and scalable forecasting methods are presented for Gaussian data. The first of these is an exact Bayesian method based on the separable spatio-temporal model discussed in Chapter 7. In this case, we evaluate the exact forecast estimates from the exact forecasting distribution and illustrate with the nyspitme data set. As an alternative to the exact method, we illustrate with the built-in forecasting method in the spTimer package. We also discuss several forecasting relevant statistics and diagrams such as the hit and false alarm rates, sharpness diagram, probability integral transform diagram, and a marginal calibration plot. We compare the forecasting methods using daily ozone pollution data in the eastern United States for a 2-week period in July 2010.
