Prediction Adjustments for Asymmetric Quadratic Loss with a Gaussian Process

Abstract

This paper presents a table of values of an adjustment factor to be used in predicting future values of a Gaussian process in the presence of asymmetric quadratic loss and which are helpful in the control of such a process. It is shown that in this case the optimal prediction is merely an additive adjustment to the conditional mean, the adjustment being the product of the standard deviation and the appropriate tabulated value.

Keywords:

• asymmetric quadratic loss
• gaussian process
• prediction