This project is concerned with studying methods of prediction based on the likelihood function alone. Determining the extent to which these predictive likelihoods may be sensibly used and the extent to which they may be inerpreted as pivotals is an important theoretical objective of this project. The close connection between these likelihoods and the parametric likelihoods allows for a study of the interrelationship between predictivism and estimation from a likelihood perspective. Prediction or forecasting is useful and important in all areas of science and is often the underlying reason for formulating and fitting statistical models in the first place. This research addresses the question of how a predictive likelihood should be extracted from an assumed statistical model.
