Computational efficient Variational Bayesian Gaussian Mixture Models via Coreset

Min Zhang, Yinlin Fu, Kevin M. Bennett, Teresa Wu

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Variational Bayesian Gaussian Mixture Model is a popular clustering algorithm with a reliable performance. However, it is noted that the model fitting process takes long time, especially when dealing with large scale data, since it utilizes the whole dataset. To address this issue, in paper we propose a new algorithm termed a weighted VBGMM via Coreset. Specifically, a new coreset construction method is first proposed to sample the data which is used to fit the model. To evaluate the algorithm, two datasets are used: 1) six rat kidney images datasets 2) three human kidney images datasets. The results show that our proposed algorithm is much faster (∼ 20 times) comparing to classic VBGMM while maintaining the similar performance on whole dataset.

Original languageEnglish
Title of host publicationIEEE CITS 2016 - 2016 International Conference on Computer, Information and Telecommunication Systems
EditorsFei Gao, Zan Li, Daniel Cascado Caballero, Jing Fan, Mohammad S. Obaidat, Petros Nicoploitidis, Kuei Fang Hsiao
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781509034406
DOIs
StatePublished - Aug 16 2016
Event2016 International Conference on Computer, Information and Telecommunication Systems, CITS 2016 - Kunming, China
Duration: Jul 6 2016Jul 8 2016

Publication series

NameIEEE CITS 2016 - 2016 International Conference on Computer, Information and Telecommunication Systems

Conference

Conference2016 International Conference on Computer, Information and Telecommunication Systems, CITS 2016
Country/TerritoryChina
CityKunming
Period07/6/1607/8/16

Keywords

  • Variational Bayesian Gaussian Mixture Model (VBGMM)
  • coreset

Fingerprint

Dive into the research topics of 'Computational efficient Variational Bayesian Gaussian Mixture Models via Coreset'. Together they form a unique fingerprint.

Cite this