A parallel EM algorithm for model-based clustering applied to the exploration of large spatio-temporal data

Wei Chen Chen; George Ostrouchov; David Pugmire; Prabhat; Michael Wehner

doi:10.1080/00401706.2013.826146

A parallel EM algorithm for model-based clustering applied to the exploration of large spatio-temporal data

Wei Chen Chen, George Ostrouchov, David Pugmire, Prabhat, Michael Wehner

Visualization

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

11 Scopus citations

Abstract

We develop a parallel expectation-maximization (EM) algorithm for multivariate Gaussian mixture models and use it to perform model-based clustering of a large climate dataset. Three variants of the EM algorithm are reformulated in parallel and a new variant that is faster is presented. All are implemented using the single program, multiple data programming model, which is able to take advantage of the combined collective memory of large distributed computer architectures to process larger datasets. Displays of the estimated mixture model rather than the data allow us to explore multivariate relationships in a way that scales to arbitrary size data. We study the performance of our methodology on simulated data and apply our methodology to a high-resolution climate dataset produced by the community atmosphere model (CAM5). This article has supplementary material online.

Original language	English
Pages (from-to)	513-523
Number of pages	11
Journal	Technometrics
Volume	55
Issue number	4
DOIs	https://doi.org/10.1080/00401706.2013.826146
State	Published - Nov 1 2013

Funding

We sincerely thank the editor, an associate editor and two reviewers for providing many insightful comments and suggestions which substantially improved this article. Work at LBNL was supported by the Regional and Global Climate Modeling Program of the Office of Biological and Environmental Research in the Department of Energy Office of Science under contract number DE-AC02-05CH11231. This research also used resources of the Oak Ridge Leadership Computing Facility at the Oak Ridge National Laboratory, which is supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC05-00OR22725.

Funders	Funder number
Office of Biological and Environmental Research in the Department of Energy Office of Science	DE-AC02-05CH11231
U.S. Department of Energy	DE-AC05-00OR22725
Office of Science

Keywords

Parallel computing
Parallel coordinate plot
Spatial time series
Unsupervised learning

Access to Document

10.1080/00401706.2013.826146

Cite this

@article{fd4db30d3c8d4cc9954dcd2f22be5ddc,

title = "A parallel EM algorithm for model-based clustering applied to the exploration of large spatio-temporal data",

abstract = "We develop a parallel expectation-maximization (EM) algorithm for multivariate Gaussian mixture models and use it to perform model-based clustering of a large climate dataset. Three variants of the EM algorithm are reformulated in parallel and a new variant that is faster is presented. All are implemented using the single program, multiple data programming model, which is able to take advantage of the combined collective memory of large distributed computer architectures to process larger datasets. Displays of the estimated mixture model rather than the data allow us to explore multivariate relationships in a way that scales to arbitrary size data. We study the performance of our methodology on simulated data and apply our methodology to a high-resolution climate dataset produced by the community atmosphere model (CAM5). This article has supplementary material online.",

keywords = "Parallel computing, Parallel coordinate plot, Spatial time series, Unsupervised learning",

author = "Chen, {Wei Chen} and George Ostrouchov and David Pugmire and Prabhat and Michael Wehner",

year = "2013",

month = nov,

day = "1",

doi = "10.1080/00401706.2013.826146",

language = "English",

volume = "55",

pages = "513--523",

journal = "Technometrics",

issn = "0040-1706",

publisher = "Taylor and Francis Group",

number = "4",

}

TY - JOUR

T1 - A parallel EM algorithm for model-based clustering applied to the exploration of large spatio-temporal data

AU - Chen, Wei Chen

AU - Ostrouchov, George

AU - Pugmire, David

AU - Prabhat,

AU - Wehner, Michael

PY - 2013/11/1

Y1 - 2013/11/1

N2 - We develop a parallel expectation-maximization (EM) algorithm for multivariate Gaussian mixture models and use it to perform model-based clustering of a large climate dataset. Three variants of the EM algorithm are reformulated in parallel and a new variant that is faster is presented. All are implemented using the single program, multiple data programming model, which is able to take advantage of the combined collective memory of large distributed computer architectures to process larger datasets. Displays of the estimated mixture model rather than the data allow us to explore multivariate relationships in a way that scales to arbitrary size data. We study the performance of our methodology on simulated data and apply our methodology to a high-resolution climate dataset produced by the community atmosphere model (CAM5). This article has supplementary material online.

AB - We develop a parallel expectation-maximization (EM) algorithm for multivariate Gaussian mixture models and use it to perform model-based clustering of a large climate dataset. Three variants of the EM algorithm are reformulated in parallel and a new variant that is faster is presented. All are implemented using the single program, multiple data programming model, which is able to take advantage of the combined collective memory of large distributed computer architectures to process larger datasets. Displays of the estimated mixture model rather than the data allow us to explore multivariate relationships in a way that scales to arbitrary size data. We study the performance of our methodology on simulated data and apply our methodology to a high-resolution climate dataset produced by the community atmosphere model (CAM5). This article has supplementary material online.

KW - Parallel computing

KW - Parallel coordinate plot

KW - Spatial time series

KW - Unsupervised learning

UR - http://www.scopus.com/inward/record.url?scp=84890031342&partnerID=8YFLogxK

U2 - 10.1080/00401706.2013.826146

DO - 10.1080/00401706.2013.826146

M3 - Article

AN - SCOPUS:84890031342

SN - 0040-1706

VL - 55

SP - 513

EP - 523

JO - Technometrics

JF - Technometrics

IS - 4

ER -

A parallel EM algorithm for model-based clustering applied to the exploration of large spatio-temporal data

Abstract

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this