Efficient Computation of Bayesian Optimal Discriminating Designs

摘要: ABSTRACTAn efficient algorithm for the determination of Bayesian optimal discriminating designs competing regression models is developed, where main focus on with general distributional assumptions beyond “classical” case normally distributed homoscedastic errors. For this purpose, we consider a version Kullback–Leibler (KL). Discretizing prior distribution leads to local KL-optimal design problems large number models. All currently available methods either require amount computation time or fail calculate design, because they can only deal efficiently few model comparisons. In article, develop new respect criterion. It demonstrated that able reasonable accuracy and computational in situa...