Title: Bayesian optimal designs via MCMC simulations: A case study in the technological field
Authors: Rossella Berni - University of Florence (Italy) [presenting]
Abstract: Optimal design criteria have recently received growing attention, both theoretically and computationally, also due to the increase of computational power. Since the 70s, there has been a long list of seminal papers about $D$ and $T$ optimality, both to estimate model parameters and to discriminate among models. Furthermore, optimal designs have been improved in a Bayesian framework, by introducing prior distributions on models and parameters and by selecting the optimal design according to the definition of an utility function and its maximization, in a decision analysis framework. Despite the generality achieved, in actual applications further flexibility is often needed: for example, when defining a utility function in which the cost of each observation depends on the value of the independent variables; also, the relevance for costs may be also evaluated by specific weights, which take environmental conditions and technological information into account. We improve optimal designs in the technological field by applying Markov Chain Monte Carlo simulations, and by evaluating: i) a hierarchical structure of the observed data; ii) an utility function including costs and weights; iii) model discrimination.