An Adaptive Derivative Free Method for Bayesian Posterior Approximation

Matti Raitoharju; Simo Ali-Löytty

doi:10.1109/LSP.2011.2179800

An Adaptive Derivative Free Method for Bayesian Posterior Approximation

Matti Raitoharju, Simo Ali-Löytty

Research output: Contribution to journal › Article › Scientific › peer-review

13 Citations (Scopus)

44 Downloads (Pure)

Abstract

In the Gaussian mixture approach a Bayesian posterior probability distribution function is approximated using a weighted sum of Gaussians. This work presents a novel method for generating a Gaussian mixture by splitting the prior taking the direction of maximum nonlinearity into account. The proposed method is computationally feasible and does not require analytical differentiation. Tests show that the method approximates the posterior better with fewer Gaussian components than existing methods.

Translated title of the contribution	An Adaptive Derivative Free Method for Bayesian Posterior Approximation
Original language	English
Article number	12436433
Pages (from-to)	87-90
Journal	IEEE Signal Processing Letters
Volume	19
Issue number	2
DOIs	https://doi.org/10.1109/LSP.2011.2179800
Publication status	Published - 2012
Publication type	A1 Journal article-refereed

Publication forum classification

Publication forum level 2

Access to Document

10.1109/LSP.2011.2179800

Raitoharju ali-loytty an adaptive derivative free method
© 2012 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

http://urn.fi/URN:NBN:fi:tty-201205311164

Cite this

@article{a9fac750771b4aab8b5b5c0073f13e02,

title = "An Adaptive Derivative Free Method for Bayesian Posterior Approximation",

abstract = "In the Gaussian mixture approach a Bayesian posterior probability distribution function is approximated using a weighted sum of Gaussians. This work presents a novel method for generating a Gaussian mixture by splitting the prior taking the direction of maximum nonlinearity into account. The proposed method is computationally feasible and does not require analytical differentiation. Tests show that the method approximates the posterior better with fewer Gaussian components than existing methods.",

author = "Matti Raitoharju and Simo Ali-L{\"o}ytty",

note = "Contribution: organisation=mat,FACT1=1<br/>Publisher name: Institute of Electrical and Electronics Engineers IEEE",

year = "2012",

doi = "10.1109/LSP.2011.2179800",

language = "English",

volume = "19",

pages = "87--90",

journal = "IEEE Signal Processing Letters",

issn = "1070-9908",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "2",

}

TY - JOUR

T1 - An Adaptive Derivative Free Method for Bayesian Posterior Approximation

AU - Raitoharju, Matti

AU - Ali-Löytty, Simo

N1 - Contribution: organisation=mat,FACT1=1<br/>Publisher name: Institute of Electrical and Electronics Engineers IEEE

PY - 2012

Y1 - 2012

N2 - In the Gaussian mixture approach a Bayesian posterior probability distribution function is approximated using a weighted sum of Gaussians. This work presents a novel method for generating a Gaussian mixture by splitting the prior taking the direction of maximum nonlinearity into account. The proposed method is computationally feasible and does not require analytical differentiation. Tests show that the method approximates the posterior better with fewer Gaussian components than existing methods.

AB - In the Gaussian mixture approach a Bayesian posterior probability distribution function is approximated using a weighted sum of Gaussians. This work presents a novel method for generating a Gaussian mixture by splitting the prior taking the direction of maximum nonlinearity into account. The proposed method is computationally feasible and does not require analytical differentiation. Tests show that the method approximates the posterior better with fewer Gaussian components than existing methods.

U2 - 10.1109/LSP.2011.2179800

DO - 10.1109/LSP.2011.2179800

M3 - Article

SN - 1070-9908

VL - 19

SP - 87

EP - 90

JO - IEEE Signal Processing Letters

JF - IEEE Signal Processing Letters

IS - 2

M1 - 12436433

ER -

An Adaptive Derivative Free Method for Bayesian Posterior Approximation

Abstract

Publication forum classification

Access to Document

Fingerprint

Cite this