Nonparametric likelihood approaches and their applications

Abstract

Empirical likelihood, first introduced by Owen (1988, 1990), is a nonparametric method in statistical inference. It is a powerful technique, which has many advantages over some other alternative methods such as normal approximation and the bootstrap. In this thesis, I shall apply the empirical likelihood to two problems. The first one is concerned with the ratio of two population means and the second is on statistical inference with possible zero observations for both one- and two-sample problems.

Date of Award	2004
Original language	English
Awarding Institution	The Hong Kong University of Science and Technology