Search

The article discusses a study in which the authors built, via computer models, designs with good prediction accuracy. The authors studied an asymptotic form of the Gussian correlation function for the stochastic process where the correlation tends to unity. After illustrating that the limiting best linear unbiased predictor involves Lagrange interpolating polynomials, they developed optimal designs based on reducing the limiting, integrated mean squared error of prediction. They then illustrated through examples how these designs make good predictors.