Topics in Model Building. Part III. Posterior Probabilities of Candidate Models in Model Discrimination.
WISCONSIN UNIV MADISON DEPT OF STATISTICS
Pagination or Media Count:
Procedures are discussed for obtaining the posterior probability for each of the candidate models in model discrimination when a prior state of indifference concerning the unknown parameter exists. The first argument makes use of information theory and arrives at the result of Box and Henson. Two additional ways of reasoning are introduced for the special case where the number of unknown parameters is identical in all the models. It is shown that these additional approaches both lead to results which are similar to the original result. Author
- Statistics and Probability