Bayesian Inference for Complex Computer Models
One of the big success stories of Bayesian inference is inference in large complex and highly structured models. A typical example is inference for computer models. Scientists use complex computer models to study the behavior of complex physical processes such as weather forecasting, disease dynamics, hydrology, traffic models, etc. Inference involves three related models, the true system, the complex simulation model and possibly a computationally more efficient emulation model. Appropriate propagation of uncertainties, good choice of emulation models, and calibration of parameters for the emulation model pose challenging inference problems reviewed in this chapter.
KeywordsComputer Model Posterior Distribution Markov Chain Monte Carlo Bayesian Inference Gaussian Process
Unable to display preview. Download preview PDF.