Finite-sample inference with monotone incomplete multivariate normal data, I

Wan Ying Chang, Donald Richards

Research output: Contribution to journalArticle

26 Citations (Scopus)

Abstract

We consider problems in finite-sample inference with two-step, monotone incomplete data drawn from Nd (μ, Σ), a multivariate normal population with mean μ and covariance matrix Σ. We derive a stochastic representation for the exact distribution of over(μ, ̂), the maximum likelihood estimator of μ. We obtain ellipsoidal confidence regions for μ through T2, a generalization of Hotelling's statistic. We derive the asymptotic distribution of, and probability inequalities for, T2 under various assumptions on the sizes of the complete and incomplete samples. Further, we establish an upper bound for the supremum distance between the probability density functions of over(μ, ̂) and over(μ, ̃), a normal approximation to over(μ, ̂).

Original languageEnglish (US)
Pages (from-to)1883-1899
Number of pages17
JournalJournal of Multivariate Analysis
Volume100
Issue number9
DOIs
StatePublished - Oct 1 2009

Fingerprint

Multivariate Normal
Covariance matrix
Probability density function
Maximum likelihood
Monotone
Statistics
Probability Inequalities
Stochastic Representation
Normal Population
Normal Approximation
Confidence Region
Exact Distribution
Incomplete Data
Supremum
Maximum Likelihood Estimator
Asymptotic distribution
Statistic
Upper bound
Hotelling
Finite sample

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Numerical Analysis
  • Statistics, Probability and Uncertainty

Cite this

@article{3a5eb16d20264687a7296fd667f741ef,
title = "Finite-sample inference with monotone incomplete multivariate normal data, I",
abstract = "We consider problems in finite-sample inference with two-step, monotone incomplete data drawn from Nd (μ, Σ), a multivariate normal population with mean μ and covariance matrix Σ. We derive a stochastic representation for the exact distribution of over(μ, ̂), the maximum likelihood estimator of μ. We obtain ellipsoidal confidence regions for μ through T2, a generalization of Hotelling's statistic. We derive the asymptotic distribution of, and probability inequalities for, T2 under various assumptions on the sizes of the complete and incomplete samples. Further, we establish an upper bound for the supremum distance between the probability density functions of over(μ, ̂) and over(μ, ̃), a normal approximation to over(μ, ̂).",
author = "Chang, {Wan Ying} and Donald Richards",
year = "2009",
month = "10",
day = "1",
doi = "10.1016/j.jmva.2009.05.003",
language = "English (US)",
volume = "100",
pages = "1883--1899",
journal = "Journal of Multivariate Analysis",
issn = "0047-259X",
publisher = "Academic Press Inc.",
number = "9",

}

Finite-sample inference with monotone incomplete multivariate normal data, I. / Chang, Wan Ying; Richards, Donald.

In: Journal of Multivariate Analysis, Vol. 100, No. 9, 01.10.2009, p. 1883-1899.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Finite-sample inference with monotone incomplete multivariate normal data, I

AU - Chang, Wan Ying

AU - Richards, Donald

PY - 2009/10/1

Y1 - 2009/10/1

N2 - We consider problems in finite-sample inference with two-step, monotone incomplete data drawn from Nd (μ, Σ), a multivariate normal population with mean μ and covariance matrix Σ. We derive a stochastic representation for the exact distribution of over(μ, ̂), the maximum likelihood estimator of μ. We obtain ellipsoidal confidence regions for μ through T2, a generalization of Hotelling's statistic. We derive the asymptotic distribution of, and probability inequalities for, T2 under various assumptions on the sizes of the complete and incomplete samples. Further, we establish an upper bound for the supremum distance between the probability density functions of over(μ, ̂) and over(μ, ̃), a normal approximation to over(μ, ̂).

AB - We consider problems in finite-sample inference with two-step, monotone incomplete data drawn from Nd (μ, Σ), a multivariate normal population with mean μ and covariance matrix Σ. We derive a stochastic representation for the exact distribution of over(μ, ̂), the maximum likelihood estimator of μ. We obtain ellipsoidal confidence regions for μ through T2, a generalization of Hotelling's statistic. We derive the asymptotic distribution of, and probability inequalities for, T2 under various assumptions on the sizes of the complete and incomplete samples. Further, we establish an upper bound for the supremum distance between the probability density functions of over(μ, ̂) and over(μ, ̃), a normal approximation to over(μ, ̂).

UR - http://www.scopus.com/inward/record.url?scp=68949204291&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=68949204291&partnerID=8YFLogxK

U2 - 10.1016/j.jmva.2009.05.003

DO - 10.1016/j.jmva.2009.05.003

M3 - Article

VL - 100

SP - 1883

EP - 1899

JO - Journal of Multivariate Analysis

JF - Journal of Multivariate Analysis

SN - 0047-259X

IS - 9

ER -