On the asymptotic distribution of the likelihood ratio statistic

被引:1
|
作者
Scott, W. F. [1 ]
机构
[1] Univ Edinburgh, Sch Math, Dept Math Sci, Edinburgh EH9 3J2, Midlothian, Scotland
关键词
contingency tables; Fisher's information; maximum likelihood estimation;
D O I
10.1080/03610920600974229
中图分类号
O21 [概率论与数理统计]; C8 [统计学];
学科分类号
020208 ; 070103 ; 0714 ;
摘要
Likelihood ratio tests are among the most frequently used in practical statistics. Important references on this subject include Wilks (1962), Rao (1965, pp. 347-352), and Agresti (1990, pp. 434-435). We give proofs of some well-known results on the limiting distribution of D = -2 log(lambda), where lambda is the likelihood ratio, under certain clearly-stated assumptions. We illustrate the theory by means of an example on contingency tables. Although results of this type are frequently used, the underlying theory is not always fully explained.
引用
收藏
页码:273 / 281
页数:9
相关论文
共 50 条