EXISTENCE AND CONSISTENCY OF MAXIMUM-LIKELIHOOD IN UPGRADED MIXTURE-MODELS

被引：16

作者：

VANDERVAART, AW ^{[1
]}

WELLNER, JA ^{[1
]}

机构：

[1] UNIV WASHINGTON,DEPT STAT GN22,SEATTLE,WA 98195

来源：

JOURNAL OF MULTIVARIATE ANALYSIS | 1992年 / 43卷 / 01期

基金：

美国国家科学基金会;

关键词：

MIXTURE MODEL; MAXIMUM LIKELIHOOD; CONSISTENCY;

D O I：

10.1016/0047-259X(92)90113-T

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：

Suppose one observes a sample of size m from the mixture density ∫ p(x|z) dη(z) and a sample of size n from the distribution η. The kernel p(x|z) is known. We show existence of the maximum likelihood estimator for η, characterize its support, and prove consistency as m, n → ∞. © 1992.

引用

页码：133 / 146

页数：14

共 16 条

[1]

Bauer H., 1981, PROBABILITY THEORY E

[2]

BHANJA J, 1988, EFFICIENT ESTIMATION

[3]

BICKEL PJ, 1992, EFFICIENT ADAPTIVE I

[4]

DUDLEY RM, 1989, REAL ANAL PROBABILIT