This article presents a Bayesian model approach for predicting a response that is a set of observed values of a function. It is assumed that there are controllable factors and randomly varying noise factors. The approach incorporates the uncertainty of the model parameters in the optimization phase. The approach is illustrated with real examples from the literature and with simulated data. Practical aspects of model building and diagnostics are also discussed.
© 2001-2024 Fundación Dialnet · Todos los derechos reservados