The estimation of normal mixtures with latent variables

Gideon Magnus, Jan R. Magnus

Research output: Contribution to JournalArticleAcademicpeer-review

Abstract

This paper considers the class of normal latent factor mixture models. It presents a method for estimating the posterior distribution of the parameters, derives analytical expressions for both the first and second derivatives of the posterior kernel (the score and Hessian), and provides posterior approximations that can be computed relatively quickly.

Original languageEnglish
Pages (from-to)1255-1269
Number of pages15
JournalCommunications in Statistics - Theory and Methods
Volume48
Issue number5
DOIs
Publication statusPublished - 8 Feb 2019

Fingerprint

Normal Mixture
Latent Variables
Factor Models
Second derivative
Posterior distribution
Mixture Model
kernel
Approximation
Class

Keywords

  • Factor analyzers
  • Hessian matrix.
  • Mixture modelling

Cite this

@article{609679ad361d414686c24f54d86649c3,
title = "The estimation of normal mixtures with latent variables",
abstract = "This paper considers the class of normal latent factor mixture models. It presents a method for estimating the posterior distribution of the parameters, derives analytical expressions for both the first and second derivatives of the posterior kernel (the score and Hessian), and provides posterior approximations that can be computed relatively quickly.",
keywords = "Factor analyzers, Hessian matrix., Mixture modelling",
author = "Gideon Magnus and Magnus, {Jan R.}",
year = "2019",
month = "2",
day = "8",
doi = "10.1080/03610926.2018.1429625",
language = "English",
volume = "48",
pages = "1255--1269",
journal = "Communications in Statistics, Theory and Methods",
issn = "0361-0926",
publisher = "Taylor and Francis Ltd.",
number = "5",

}

The estimation of normal mixtures with latent variables. / Magnus, Gideon; Magnus, Jan R.

In: Communications in Statistics - Theory and Methods, Vol. 48, No. 5, 08.02.2019, p. 1255-1269.

Research output: Contribution to JournalArticleAcademicpeer-review

TY - JOUR

T1 - The estimation of normal mixtures with latent variables

AU - Magnus, Gideon

AU - Magnus, Jan R.

PY - 2019/2/8

Y1 - 2019/2/8

N2 - This paper considers the class of normal latent factor mixture models. It presents a method for estimating the posterior distribution of the parameters, derives analytical expressions for both the first and second derivatives of the posterior kernel (the score and Hessian), and provides posterior approximations that can be computed relatively quickly.

AB - This paper considers the class of normal latent factor mixture models. It presents a method for estimating the posterior distribution of the parameters, derives analytical expressions for both the first and second derivatives of the posterior kernel (the score and Hessian), and provides posterior approximations that can be computed relatively quickly.

KW - Factor analyzers

KW - Hessian matrix.

KW - Mixture modelling

UR - http://www.scopus.com/inward/record.url?scp=85041501537&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85041501537&partnerID=8YFLogxK

U2 - 10.1080/03610926.2018.1429625

DO - 10.1080/03610926.2018.1429625

M3 - Article

VL - 48

SP - 1255

EP - 1269

JO - Communications in Statistics, Theory and Methods

JF - Communications in Statistics, Theory and Methods

SN - 0361-0926

IS - 5

ER -