COFADMM

A computational features selection with alternating Direction Method of Multipliers

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

Due to the explosion in size and complexity of Big Data, it is increasingly important to be able to solve problems with very large number of features. Classical feature selection procedures involves combinatorial optimization, with computational time increasing exponentially with the number of features. During the last decade, penalized regression has emerged as an attractive alternative for regularization and high dimensional feature selection problems. Alternating Direction Method of Multipliers (ADMM) optimization is suited for distributed convex optimization and distributed computing for big data. The purpose of this paper is to propose a broader algorithm COFADMM which combines the strength of convex penalized techniques in feature selection for big data and the power of the ADMM for optimization. We show that combining the ADMM algorithm with COFADMM can provide a path of solutions efficiently and quickly. COFADMM is easy to use, is available in C, Matlab upon request from the corresponding author.

Original languageEnglish
Title of host publicationProcedia Computer Science
PublisherElsevier
Pages821-830
Number of pages10
Volume29
DOIs
Publication statusPublished - 2014
Event14th Annual International Conference on Computational Science, ICCS 2014 - Cairns, QLD, Australia
Duration: 10 Jun 201412 Jun 2014

Other

Other14th Annual International Conference on Computational Science, ICCS 2014
CountryAustralia
CityCairns, QLD
Period10/6/1412/6/14

Fingerprint

Feature extraction
Convex optimization
Combinatorial optimization
Distributed computer systems
Explosions
Big data

Keywords

  • ADMM algorithm
  • Coordinate descent algorithm
  • Features selection
  • Lasso
  • Least angle regression algorithm

ASJC Scopus subject areas

  • Computer Science(all)

Cite this

COFADMM : A computational features selection with alternating Direction Method of Multipliers. / El Anbari, Mohammed; Alam, Sidra; Bensmail, Halima.

Procedia Computer Science. Vol. 29 Elsevier, 2014. p. 821-830.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

El Anbari, M, Alam, S & Bensmail, H 2014, COFADMM: A computational features selection with alternating Direction Method of Multipliers. in Procedia Computer Science. vol. 29, Elsevier, pp. 821-830, 14th Annual International Conference on Computational Science, ICCS 2014, Cairns, QLD, Australia, 10/6/14. https://doi.org/10.1016/j.procs.2014.05.074
@inproceedings{367cf81dfa29427db689996267d6797f,
title = "COFADMM: A computational features selection with alternating Direction Method of Multipliers",
abstract = "Due to the explosion in size and complexity of Big Data, it is increasingly important to be able to solve problems with very large number of features. Classical feature selection procedures involves combinatorial optimization, with computational time increasing exponentially with the number of features. During the last decade, penalized regression has emerged as an attractive alternative for regularization and high dimensional feature selection problems. Alternating Direction Method of Multipliers (ADMM) optimization is suited for distributed convex optimization and distributed computing for big data. The purpose of this paper is to propose a broader algorithm COFADMM which combines the strength of convex penalized techniques in feature selection for big data and the power of the ADMM for optimization. We show that combining the ADMM algorithm with COFADMM can provide a path of solutions efficiently and quickly. COFADMM is easy to use, is available in C, Matlab upon request from the corresponding author.",
keywords = "ADMM algorithm, Coordinate descent algorithm, Features selection, Lasso, Least angle regression algorithm",
author = "{El Anbari}, Mohammed and Sidra Alam and Halima Bensmail",
year = "2014",
doi = "10.1016/j.procs.2014.05.074",
language = "English",
volume = "29",
pages = "821--830",
booktitle = "Procedia Computer Science",
publisher = "Elsevier",

}

TY - GEN

T1 - COFADMM

T2 - A computational features selection with alternating Direction Method of Multipliers

AU - El Anbari, Mohammed

AU - Alam, Sidra

AU - Bensmail, Halima

PY - 2014

Y1 - 2014

N2 - Due to the explosion in size and complexity of Big Data, it is increasingly important to be able to solve problems with very large number of features. Classical feature selection procedures involves combinatorial optimization, with computational time increasing exponentially with the number of features. During the last decade, penalized regression has emerged as an attractive alternative for regularization and high dimensional feature selection problems. Alternating Direction Method of Multipliers (ADMM) optimization is suited for distributed convex optimization and distributed computing for big data. The purpose of this paper is to propose a broader algorithm COFADMM which combines the strength of convex penalized techniques in feature selection for big data and the power of the ADMM for optimization. We show that combining the ADMM algorithm with COFADMM can provide a path of solutions efficiently and quickly. COFADMM is easy to use, is available in C, Matlab upon request from the corresponding author.

AB - Due to the explosion in size and complexity of Big Data, it is increasingly important to be able to solve problems with very large number of features. Classical feature selection procedures involves combinatorial optimization, with computational time increasing exponentially with the number of features. During the last decade, penalized regression has emerged as an attractive alternative for regularization and high dimensional feature selection problems. Alternating Direction Method of Multipliers (ADMM) optimization is suited for distributed convex optimization and distributed computing for big data. The purpose of this paper is to propose a broader algorithm COFADMM which combines the strength of convex penalized techniques in feature selection for big data and the power of the ADMM for optimization. We show that combining the ADMM algorithm with COFADMM can provide a path of solutions efficiently and quickly. COFADMM is easy to use, is available in C, Matlab upon request from the corresponding author.

KW - ADMM algorithm

KW - Coordinate descent algorithm

KW - Features selection

KW - Lasso

KW - Least angle regression algorithm

UR - http://www.scopus.com/inward/record.url?scp=84902771967&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84902771967&partnerID=8YFLogxK

U2 - 10.1016/j.procs.2014.05.074

DO - 10.1016/j.procs.2014.05.074

M3 - Conference contribution

VL - 29

SP - 821

EP - 830

BT - Procedia Computer Science

PB - Elsevier

ER -