General Bayesian Updating
Stephen Walker (University of Kent)
Friday 27th January, 2012 15:00-16:00 Maths 203
All unknown quantities of interest can have prior distributions reflecting beliefs as to their ultimate location. But not all connections between observations and parameters of interest are based on probability models. When observations and parameters are cinnected via loss functions a means by which to update the prior is needed. This talk will discuss how this can be done in a coherent way.