Acceleration of the EM algorithm using the Vector Aitken method and its Steffensen form

Xu Guo, Qiu yue Li, Wang li Xu

Research output: Contribution to journalArticle

Abstract

Based on Vector Aitken (VA) method, we propose an acceleration Expectation-Maximization (EM) algorithm, VA-accelerated EM algorithm, whose convergence speed is faster than that of EM algorithm. The VA-accelerated EM algorithm does not use the information matrix but only uses the sequence of estimates obtained from iterations of the EM algorithm, thus it keeps the flexibility and simplicity of the EM algorithm. Considering Steffensen iterative process, we have also given the Steffensen form of the VA-accelerated EM algorithm. It can be proved that the reform process is quadratic convergence. Numerical analysis illustrate the proposed methods are efficient and faster than EM algorithm.

Original languageEnglish (US)
Pages (from-to)175-182
Number of pages8
JournalActa Mathematicae Applicatae Sinica
Volume33
Issue number1
DOIs
StatePublished - Feb 1 2017

Fingerprint

Expectation-maximization Algorithm
Quadratic Convergence
Information Matrix
Form
Convergence Speed
Iterative Process
Numerical analysis
Numerical Analysis
Simplicity
Flexibility
Iteration
Estimate

All Science Journal Classification (ASJC) codes

  • Applied Mathematics

Cite this

@article{75eb7a67877545d7807821735848958d,
title = "Acceleration of the EM algorithm using the Vector Aitken method and its Steffensen form",
abstract = "Based on Vector Aitken (VA) method, we propose an acceleration Expectation-Maximization (EM) algorithm, VA-accelerated EM algorithm, whose convergence speed is faster than that of EM algorithm. The VA-accelerated EM algorithm does not use the information matrix but only uses the sequence of estimates obtained from iterations of the EM algorithm, thus it keeps the flexibility and simplicity of the EM algorithm. Considering Steffensen iterative process, we have also given the Steffensen form of the VA-accelerated EM algorithm. It can be proved that the reform process is quadratic convergence. Numerical analysis illustrate the proposed methods are efficient and faster than EM algorithm.",
author = "Xu Guo and Li, {Qiu yue} and Xu, {Wang li}",
year = "2017",
month = "2",
day = "1",
doi = "10.1007/s10255-017-0648-3",
language = "English (US)",
volume = "33",
pages = "175--182",
journal = "Acta Mathematicae Applicatae Sinica",
issn = "0168-9673",
publisher = "Springer Verlag",
number = "1",

}

Acceleration of the EM algorithm using the Vector Aitken method and its Steffensen form. / Guo, Xu; Li, Qiu yue; Xu, Wang li.

In: Acta Mathematicae Applicatae Sinica, Vol. 33, No. 1, 01.02.2017, p. 175-182.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Acceleration of the EM algorithm using the Vector Aitken method and its Steffensen form

AU - Guo, Xu

AU - Li, Qiu yue

AU - Xu, Wang li

PY - 2017/2/1

Y1 - 2017/2/1

N2 - Based on Vector Aitken (VA) method, we propose an acceleration Expectation-Maximization (EM) algorithm, VA-accelerated EM algorithm, whose convergence speed is faster than that of EM algorithm. The VA-accelerated EM algorithm does not use the information matrix but only uses the sequence of estimates obtained from iterations of the EM algorithm, thus it keeps the flexibility and simplicity of the EM algorithm. Considering Steffensen iterative process, we have also given the Steffensen form of the VA-accelerated EM algorithm. It can be proved that the reform process is quadratic convergence. Numerical analysis illustrate the proposed methods are efficient and faster than EM algorithm.

AB - Based on Vector Aitken (VA) method, we propose an acceleration Expectation-Maximization (EM) algorithm, VA-accelerated EM algorithm, whose convergence speed is faster than that of EM algorithm. The VA-accelerated EM algorithm does not use the information matrix but only uses the sequence of estimates obtained from iterations of the EM algorithm, thus it keeps the flexibility and simplicity of the EM algorithm. Considering Steffensen iterative process, we have also given the Steffensen form of the VA-accelerated EM algorithm. It can be proved that the reform process is quadratic convergence. Numerical analysis illustrate the proposed methods are efficient and faster than EM algorithm.

UR - http://www.scopus.com/inward/record.url?scp=85015620189&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85015620189&partnerID=8YFLogxK

U2 - 10.1007/s10255-017-0648-3

DO - 10.1007/s10255-017-0648-3

M3 - Article

AN - SCOPUS:85015620189

VL - 33

SP - 175

EP - 182

JO - Acta Mathematicae Applicatae Sinica

JF - Acta Mathematicae Applicatae Sinica

SN - 0168-9673

IS - 1

ER -