Compatible prior distributions
A. P. Dawid, Steffen Lilholt Lauritzen · 2000
We investigate two approaches to constructing compatible prior laws over alter-native models: ‘projection ’ and ‘conditioning’. Each of these is shown to require additional inputs. We suggest that these can be chosen in a natural way in each case, leading to ‘Kullback-Leibler projection ’ and ‘Jeffreys conditioning’. We recommend the former for the case of coexisting models, and the latter for com-peting models.