A stable hybrid method for feature subset selection using particle swarm optimization with local search

Hassen Dhrif, Miroslav Kubat, Luis G.S. Giraldo, Stefan Wuchty

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

Abstract

The determination of a small set of biomarkers to make a diagnostic call can be formulated as a feature subset selection (FSS) problem to find a small set of genes with high relevance for the underlying classification task and low mutual redundancy. However, repeated application of a heuristic, evolutionary FSS technique usually fails to produce consistent results. Here, we introduce COMB-PSO-LS, a novel hybrid (wrapper-filter) FSS algorithm based on Particle Swarm Optimization (PSO) that features a local search strategy to select the least dependent and most relevant feature subsets. In particular, we employ a Randomized Dependence Coefficient (RDC)-based filter technique to guide the search process of the particle swarm, allowing the selection of highly relevant and consistent features. Classifying cancer samples through patient gene expression profiles, we found that COMB-PSO-LS provides highly stable and non-redundant gene subsets that are relevant for the classification process, outperforming standard PSO methods.

Original languageEnglish (US)
Title of host publicationGECCO 2019 - Proceedings of the 2019 Genetic and Evolutionary Computation Conference
PublisherAssociation for Computing Machinery, Inc
Pages13-21
Number of pages9
ISBN (Electronic)9781450361118
DOIs
StatePublished - Jul 13 2019
Event2019 Genetic and Evolutionary Computation Conference, GECCO 2019 - Prague, Czech Republic
Duration: Jul 13 2019Jul 17 2019

Publication series

NameGECCO 2019 - Proceedings of the 2019 Genetic and Evolutionary Computation Conference

Conference

Conference2019 Genetic and Evolutionary Computation Conference, GECCO 2019
CountryCzech Republic
CityPrague
Period7/13/197/17/19

Fingerprint

Feature Subset Selection
Hybrid Method
Set theory
Particle swarm optimization (PSO)
Local Search
Particle Swarm Optimization
Genes
Filter
Gene
Particle Swarm
Gene Expression Profile
Subset
Wrapper
Biomarkers
Search Strategy
Gene expression
Redundancy
Optimization Methods
Diagnostics
Cancer

ASJC Scopus subject areas

  • Computational Mathematics

Cite this

Dhrif, H., Kubat, M., Giraldo, L. G. S., & Wuchty, S. (2019). A stable hybrid method for feature subset selection using particle swarm optimization with local search. In GECCO 2019 - Proceedings of the 2019 Genetic and Evolutionary Computation Conference (pp. 13-21). (GECCO 2019 - Proceedings of the 2019 Genetic and Evolutionary Computation Conference). Association for Computing Machinery, Inc. https://doi.org/10.1145/3321707.3321816

A stable hybrid method for feature subset selection using particle swarm optimization with local search. / Dhrif, Hassen; Kubat, Miroslav; Giraldo, Luis G.S.; Wuchty, Stefan.

GECCO 2019 - Proceedings of the 2019 Genetic and Evolutionary Computation Conference. Association for Computing Machinery, Inc, 2019. p. 13-21 (GECCO 2019 - Proceedings of the 2019 Genetic and Evolutionary Computation Conference).

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

Dhrif, H, Kubat, M, Giraldo, LGS & Wuchty, S 2019, A stable hybrid method for feature subset selection using particle swarm optimization with local search. in GECCO 2019 - Proceedings of the 2019 Genetic and Evolutionary Computation Conference. GECCO 2019 - Proceedings of the 2019 Genetic and Evolutionary Computation Conference, Association for Computing Machinery, Inc, pp. 13-21, 2019 Genetic and Evolutionary Computation Conference, GECCO 2019, Prague, Czech Republic, 7/13/19. https://doi.org/10.1145/3321707.3321816
Dhrif H, Kubat M, Giraldo LGS, Wuchty S. A stable hybrid method for feature subset selection using particle swarm optimization with local search. In GECCO 2019 - Proceedings of the 2019 Genetic and Evolutionary Computation Conference. Association for Computing Machinery, Inc. 2019. p. 13-21. (GECCO 2019 - Proceedings of the 2019 Genetic and Evolutionary Computation Conference). https://doi.org/10.1145/3321707.3321816
Dhrif, Hassen ; Kubat, Miroslav ; Giraldo, Luis G.S. ; Wuchty, Stefan. / A stable hybrid method for feature subset selection using particle swarm optimization with local search. GECCO 2019 - Proceedings of the 2019 Genetic and Evolutionary Computation Conference. Association for Computing Machinery, Inc, 2019. pp. 13-21 (GECCO 2019 - Proceedings of the 2019 Genetic and Evolutionary Computation Conference).
@inproceedings{1728acb94b2a4422a25cd3d1a8c30873,
title = "A stable hybrid method for feature subset selection using particle swarm optimization with local search",
abstract = "The determination of a small set of biomarkers to make a diagnostic call can be formulated as a feature subset selection (FSS) problem to find a small set of genes with high relevance for the underlying classification task and low mutual redundancy. However, repeated application of a heuristic, evolutionary FSS technique usually fails to produce consistent results. Here, we introduce COMB-PSO-LS, a novel hybrid (wrapper-filter) FSS algorithm based on Particle Swarm Optimization (PSO) that features a local search strategy to select the least dependent and most relevant feature subsets. In particular, we employ a Randomized Dependence Coefficient (RDC)-based filter technique to guide the search process of the particle swarm, allowing the selection of highly relevant and consistent features. Classifying cancer samples through patient gene expression profiles, we found that COMB-PSO-LS provides highly stable and non-redundant gene subsets that are relevant for the classification process, outperforming standard PSO methods.",
author = "Hassen Dhrif and Miroslav Kubat and Giraldo, {Luis G.S.} and Stefan Wuchty",
year = "2019",
month = "7",
day = "13",
doi = "10.1145/3321707.3321816",
language = "English (US)",
series = "GECCO 2019 - Proceedings of the 2019 Genetic and Evolutionary Computation Conference",
publisher = "Association for Computing Machinery, Inc",
pages = "13--21",
booktitle = "GECCO 2019 - Proceedings of the 2019 Genetic and Evolutionary Computation Conference",

}

TY - GEN

T1 - A stable hybrid method for feature subset selection using particle swarm optimization with local search

AU - Dhrif, Hassen

AU - Kubat, Miroslav

AU - Giraldo, Luis G.S.

AU - Wuchty, Stefan

PY - 2019/7/13

Y1 - 2019/7/13

N2 - The determination of a small set of biomarkers to make a diagnostic call can be formulated as a feature subset selection (FSS) problem to find a small set of genes with high relevance for the underlying classification task and low mutual redundancy. However, repeated application of a heuristic, evolutionary FSS technique usually fails to produce consistent results. Here, we introduce COMB-PSO-LS, a novel hybrid (wrapper-filter) FSS algorithm based on Particle Swarm Optimization (PSO) that features a local search strategy to select the least dependent and most relevant feature subsets. In particular, we employ a Randomized Dependence Coefficient (RDC)-based filter technique to guide the search process of the particle swarm, allowing the selection of highly relevant and consistent features. Classifying cancer samples through patient gene expression profiles, we found that COMB-PSO-LS provides highly stable and non-redundant gene subsets that are relevant for the classification process, outperforming standard PSO methods.

AB - The determination of a small set of biomarkers to make a diagnostic call can be formulated as a feature subset selection (FSS) problem to find a small set of genes with high relevance for the underlying classification task and low mutual redundancy. However, repeated application of a heuristic, evolutionary FSS technique usually fails to produce consistent results. Here, we introduce COMB-PSO-LS, a novel hybrid (wrapper-filter) FSS algorithm based on Particle Swarm Optimization (PSO) that features a local search strategy to select the least dependent and most relevant feature subsets. In particular, we employ a Randomized Dependence Coefficient (RDC)-based filter technique to guide the search process of the particle swarm, allowing the selection of highly relevant and consistent features. Classifying cancer samples through patient gene expression profiles, we found that COMB-PSO-LS provides highly stable and non-redundant gene subsets that are relevant for the classification process, outperforming standard PSO methods.

UR - http://www.scopus.com/inward/record.url?scp=85072343407&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85072343407&partnerID=8YFLogxK

U2 - 10.1145/3321707.3321816

DO - 10.1145/3321707.3321816

M3 - Conference contribution

AN - SCOPUS:85072343407

T3 - GECCO 2019 - Proceedings of the 2019 Genetic and Evolutionary Computation Conference

SP - 13

EP - 21

BT - GECCO 2019 - Proceedings of the 2019 Genetic and Evolutionary Computation Conference

PB - Association for Computing Machinery, Inc

ER -