Smith, J. Q.

The aim of this thesis is to generalise Bayesian Forecasting processes to models where normality assumptions are, not appropriate. In particular I develop models that can change their minds and I utilise Catastrophe Theory in their description. Under squared-error loss types of …