LIMITING PROPERTIES OF LIKELIHOODS AND INFERENCE.
FLORIDA STATE UNIV TALLAHASSEE DEPT OF STATISTICS
Pagination or Media Count:
The paper supports to some extent the idea that the likelihood principle is applicable within a wide area. The sphere of applicability does not depend on the form of the sample space but on the form of likelihood functions likelihoods, for short. If the likelihoods are exactly normal, then being given a likelihood function we may draw inference ignoring the statistical model in the background. The main purpose of the paper is to investigate how far the background model may be ignored, if the likelihoods are only approximately normal. The investigation will be done in the framework of asymptotic theory, whose basic features the paper discusses. Author
- Statistics and Probability