A note on EM algorithm for mixture models

被引：28

作者：

Yao, Weixin ^{[1
]}

机构：

[1] Kansas State Univ, Dept Stat, Manhattan, KS 66506 USA

来源：

STATISTICS & PROBABILITY LETTERS | 2013年 / 83卷 / 02期

关键词：

Adaptive regression; EM algorithm; Edge-preserving smoothers; Mode; Robust regression; LIKELIHOOD; REGRESSION;

D O I：

10.1016/j.spl.2012.10.017

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：

Expectation-maximization (EM) algorithm has been used to maximize the likelihood function or posterior when the model contains unobserved latent variables. One main important application of EM algorithm is to find the maximum likelihood estimator for mixture models. In this article, we propose an EM type algorithm to maximize a class of mixture type objective functions. In addition, we prove the monotone ascending property of the proposed algorithm and discuss some of its applications. (c) 2012 Elsevier B.V. All rights reserved.

引用

页码：519 / 526

页数：8

共 50 条

[21] EM Estimation for Finite Mixture Models with Known Mixture Component Size [J].

Teel, Chen ;

Park, Taeyoung ;

Sampson, Allan R. .

COMMUNICATIONS IN STATISTICS-SIMULATION AND COMPUTATION, 2015, 44 (06) :1545-1556

[22] Analytic calculations for the EM algorithm for multivariate skew-t mixture models [J].

Vrbik, I. ;

McNicholas, P. D. .

STATISTICS & PROBABILITY LETTERS, 2012, 82 (06) :1169-1174

[23] Recursive EM algorithm for finite mixture models with application to Internet traffic modeling [J].

Liu, Z ;

Almhana, J ;

Choulakian, V ;

McGorman, R .

SECOND ANNUAL CONFERENCE ON COMMUNICATION NETWORKS AND SERVICES RESEARCH, PROCEEDINGS, 2004, :198-207

[24] An EM algorithm for the block mixture model [J].

Govaert, G ;

Nadif, M .

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2005, 27 (04) :643-647

[25] EM Algorithm for Normal Mixture Likelihoods [J].

Viwatwongkasem, Chukiat .

2018 6TH INTERNATIONAL ELECTRICAL ENGINEERING CONGRESS (IEECON), 2018,

[26] Initializing the EM Algorithm for Univariate Gaussian, Multi-Component, Heteroscedastic Mixture Models by Dynamic Programming Partitions [J].

Polanski, Andrzej ;

Marczyk, Michal ;

Pietrowska, Monika ;

Widlak, Piotr ;

Polanska, Joanna .

INTERNATIONAL JOURNAL OF COMPUTATIONAL METHODS, 2018, 15 (03)

[27] Two-layer EM algorithm for ALD mixture regression models: A new solution to composite quantile regression [J].

Wang, Shangshan ;

Xiang, Liming .

COMPUTATIONAL STATISTICS & DATA ANALYSIS, 2017, 115 :136-154

[28] A parallel EM algorithm for Gaussian Mixture Models implemented on a NUMA system using OpenMP [J].

Kwedlo, Wojciech .

2014 22ND EUROMICRO INTERNATIONAL CONFERENCE ON PARALLEL, DISTRIBUTED, AND NETWORK-BASED PROCESSING (PDP 2014), 2014, :292-298

[29] Comparing Several Methods to Fit Finite Mixture Models to Grouped Data by the EM Algorithm [J].

Wengrzik, Joanna ;

Timm, Juergen .

WORLD CONGRESS ON ENGINEERING, WCE 2011, VOL I, 2011, :329-332

[30] Asymptotic properties of the EM algorithm estimate for normal mixture models with component specific variances [J].

Nityasuddhi, D ;

Böhning, D .

COMPUTATIONAL STATISTICS & DATA ANALYSIS, 2003, 41 (3-4) :591-601

← 1 2 3 4 5 →