Optimal Experiment Design, Bayesian
In the design of experiments, we choose what observations we will make and under what conditions. In Bayesian experimental design, we regard this choice as a decision and apply the ideas of Bayesian decision analysis ( Bayesian Decision Analysis). We take into account prior beliefs ( Bayesian Inference) about relevant unknown quantities and also the various costs and benefits associated with the experiment, where the benefits are typically a result of a gain in knowledge or reduction in uncertainty. The values that we place on these costs and benefits are expressed through a utility function. We choose the design which maximizes our expectation, before the experiment, of the value of the utility function after the experiment. This does not necessarily mean that we use Bayesian inference to analyze the results of the experiment.
The design of an experiment is a choice of what observations we are going to make or, in some cases, a rule for making...
- Berry SM, Carlin BP, Lee JJ, Müller P (2010) Bayesian adaptive methods for clinical trials. Chapman & Hall/CRC Press, LondonGoogle Scholar
- Chaloner K, Verdinelli I (1995) Bayesian experimental design: a review. Stat Sci 10:273–304Google Scholar
- Kadane JB, Seidenfeld T (1990) Randomization in a Bayesian perspective. J Stat Plan Inference 25:329–345Google Scholar
- Müller P (1999) Simulation-based optimal design. In: Bernardo JM, Berger JO, Dawid AP, Smith AFM (eds) Bayesian statistics, vol 6. Oxford University Press, Oxford, pp 459–474Google Scholar
- O’Hagan A (1994) Kendall’s advanced theory of statistics, volume 2B: Bayesian inference. Edward Arnold, LondonGoogle Scholar
- Smith JQ (2010) Bayesian decision analysis: principles and practice. Cambridge University Press, CambridgeGoogle Scholar
- Spiegelhalter DJ, Abrams KR, Myles JP (2004) Bayesian approaches to clinical trials and health-care evaluation. Wiley, ChichesterGoogle Scholar