Optimality properties of constrained maximum-likelihood estimates when some parameter values lie on the boundary /

Bibliographic Details
Main Author: Hintze, Jerry Lund
Other Authors: Gates, Charles E. (degree committee member.), Ringer, Larry J. (degree committee member.)
Format: Thesis Book
Language:English
Published: [College Station, Tex.] : Hintze, 1976.
Subjects:
Online Access:Link to ProQuest copy
Link to OAKTrust copy
Description
Abstract:In this dissertation we establish asymptotic optimally properties for constrained maximum-likelihood estimates when some parameter values lie on the boundary. Specifically we show that when the parameter space is restricted to be the positive quadrant and one or two parameters are zero, the asymptotic generalized mean square error of the constrained maximum-likelihood estimate of a parameter vector is less than that of any estimate which is consistent for the whole parameter space. We prove that these results also hold when the restricted parameter space is of the type g(θ) (greater than or equal to) 0, where θ is the parameter vector and g is a continuous, differentiable, convex function. Finally, we extend these results to the case of estimating parameters from specifically censored likelihoods when some parameters are on the boundary.
Item Description:"Major subject: Statistics."
Vita.
Physical Description:v, 50 leaves ; 28 cm
Bibliography:Includes bibliographical references (leaf 46).