Fast multi-output relevance vector regression

Youngmin Ha, Hai Zhang

Research output: Contribution to journalArticle

Abstract

This paper has applied the matrix Gaussian distribution of the likelihood function of the complete data set to reduce time complexity of multi-output relevance vector regression from O(VM^3) to O(V^3 +M^3), where V and M are the number of output dimensions and basis functions respectively and V < M. Our experimental results demonstrate that the proposed method is more competitive and faster than the existing methods like Thayananthan et al. (2008). Its computational efficiency and accuracy can be attributed to the different model specifications of the likelihood of the data, as the existing method expresses the likelihood of the training data as the product of Gaussian distributions whereas the proposed method expresses it as the matrix Gaussian distribution.
LanguageEnglish
JournalEconomic Modelling
Publication statusAccepted/In press - 9 Apr 2019

Fingerprint

Model specification

Keywords

  • relevance vector regression
  • machine learning
  • sparse Bayesian learning

Cite this

@article{448bf95e533840948935551ae93f1467,
title = "Fast multi-output relevance vector regression",
abstract = "This paper has applied the matrix Gaussian distribution of the likelihood function of the complete data set to reduce time complexity of multi-output relevance vector regression from O(VM^3) to O(V^3 +M^3), where V and M are the number of output dimensions and basis functions respectively and V < M. Our experimental results demonstrate that the proposed method is more competitive and faster than the existing methods like Thayananthan et al. (2008). Its computational efficiency and accuracy can be attributed to the different model specifications of the likelihood of the data, as the existing method expresses the likelihood of the training data as the product of Gaussian distributions whereas the proposed method expresses it as the matrix Gaussian distribution.",
keywords = "relevance vector regression, machine learning, sparse Bayesian learning",
author = "Youngmin Ha and Hai Zhang",
year = "2019",
month = "4",
day = "9",
language = "English",
journal = "Economic Modelling",
issn = "0264-9993",

}

Fast multi-output relevance vector regression. / Ha, Youngmin ; Zhang, Hai.

In: Economic Modelling, 09.04.2019.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Fast multi-output relevance vector regression

AU - Ha, Youngmin

AU - Zhang, Hai

PY - 2019/4/9

Y1 - 2019/4/9

N2 - This paper has applied the matrix Gaussian distribution of the likelihood function of the complete data set to reduce time complexity of multi-output relevance vector regression from O(VM^3) to O(V^3 +M^3), where V and M are the number of output dimensions and basis functions respectively and V < M. Our experimental results demonstrate that the proposed method is more competitive and faster than the existing methods like Thayananthan et al. (2008). Its computational efficiency and accuracy can be attributed to the different model specifications of the likelihood of the data, as the existing method expresses the likelihood of the training data as the product of Gaussian distributions whereas the proposed method expresses it as the matrix Gaussian distribution.

AB - This paper has applied the matrix Gaussian distribution of the likelihood function of the complete data set to reduce time complexity of multi-output relevance vector regression from O(VM^3) to O(V^3 +M^3), where V and M are the number of output dimensions and basis functions respectively and V < M. Our experimental results demonstrate that the proposed method is more competitive and faster than the existing methods like Thayananthan et al. (2008). Its computational efficiency and accuracy can be attributed to the different model specifications of the likelihood of the data, as the existing method expresses the likelihood of the training data as the product of Gaussian distributions whereas the proposed method expresses it as the matrix Gaussian distribution.

KW - relevance vector regression

KW - machine learning

KW - sparse Bayesian learning

UR - https://www.sciencedirect.com/journal/economic-modelling

M3 - Article

JO - Economic Modelling

T2 - Economic Modelling

JF - Economic Modelling

SN - 0264-9993

ER -