Application of Likelihood Methods to Models Involving Large Numbers of Parameters

摘要: [Read before the ROYAL STATISTICAL SOCIETY at a meeting organized by RESEARCH SECTION on Wednesday, March 11th, 1970, Professor J. DURBIN in Chair] SUMMARY Likelihood methods of dealing with some multiparameter problems are introduced and exemplified. Specifically, eliminating nuisance parameters from likelihood function so that inferences can be made about interest considered. In this regard integrated likelihoods, maximum relative conditional marginal likelihoods second-order their uses illustrated examples. Marginal dependent upon factorings function. They applied to linear functional relationship related models found give intuitively appealing results. These indicate many situations commonly encountered objective unwanted adopted. This gives an alternative method interpreting offered Bayesian approach.