CMStatistics 2017: Start Registration
View Submission - CMStatistics
Title: Bayesian analysis of data from experiments subject to restricted randomisation Authors:  Sadiah Aljeddani - University of Southampton (United Kingdom) [presenting]
Kalliopi Mylona - Carlos III University of Madrid (Spain)
David Woods - University of Southampton (United Kingdom)
Abstract: Designs with restricted randomisation are often more cost-effective than completely randomised designs. They are popular in industry both for developing new products and processes and for quality improvement experiments. Model and variable selection for such experiments is crucial. We study the performance of Bayesian variable selection for analysing data from experiments with restricted randomisation using shrinkage priors that allow simultaneous model selection and parameter estimation. The properties of these procedures are illustrated via simulation studies and the analysis of exemplar data sets.