RESEARCH IN STATISTICS, INFORMATION THEORY, AND ANALYSIS.
CORNELL UNIV ITHACA N Y DEPT OF MATHEMATICS
Pagination or Media Count:
The classical theory of asymptotically efficient estimators is the theory of maximum likelihood m.l. estimators. The latter theory has the following inadequacies 1 it applies only in the so-called regular case 2 only estimators which are asymptotically normal are considered 3 the theory is limited principally to one dimensional parameters 4 the theory is limited principally to the case of independent, identically distributed observations. In papers 6, by Wolfowitz and 8,13, and 16 by Wolfowitz and Weiss there is developed a theory of asymptotically efficient estimators which has none of these inadequacies, which seems to apply to every parametric problem which appears in applications, and which includes the theory of m.l. as a special case. The theory explains why the m.l. estimator works and seems to solve the problem of asymptotically efficient estimators. The paper of this group which is easiest to read is probably no. 13. Author
- Statistics and Probability