Comparison of Feature Selection Methods for High-Dimensional Small Datasets

Harald Rietdijk; Patricia Conde-Cespedes; Talko  Dijkhuis; Hilbrand Oldenhuis; Maria Trocan

doi:10.1109/ISCAS66217.2026.11562723

Comparison of Feature Selection Methods for High-Dimensional Small Datasets

Vertaalde titel van de bijdrage: Vergelijking van methoden voor kenmerkselectie bij kleine datasets met een groot aantal dimensies

Harald Rietdijk
, Patricia Conde-Cespedes
, Talko Dijkhuis
, Hilbrand Oldenhuis
, Maria Trocan

Isep

Onderzoeksoutput: Paper › Academic

Samenvatting

Technieken op het gebied van machine learning (ML) worden op grote schaal toegepast in de gezondheidszorg, variërend van precisiegeneeskunde tot de mogelijkheid van preventieve zorg. Een algemeen probleem bij klinische studies is de combinatie van hoge dimensionaliteit en kleine steekproefomvang, wat het proces van het extraheren van significante indicatoren uit de feature-set bemoeilijkt. In dit artikel vergelijken we de kwaliteit van drie methoden voor kenmerkselectie wanneer het aantal steekproeven laag is in vergelijking met het aantal kenmerken. De resultaten worden vergeleken met die van eerder onderzoek waarin een optimalisatieproces werd toegepast op recursieve feature eliminatie. De twee filtermethoden, Anova en chi-kwadraat, konden geen stabiele resultaten opleveren, terwijl de embedded Lasso-methode veelbelovendere resultaten liet zien.

Vertaalde titel van de bijdrage	Vergelijking van methoden voor kenmerkselectie bij kleine datasets met een groot aantal dimensies
Originele taal-2	English
Pagina's	2893-2897
Aantal pagina's	5
DOI's	https://doi.org/10.1109/ISCAS66217.2026.11562723
Status	Published - 24 mei 2026
Evenement	2026 IEEE International Symposium on Circuits and Systems - International Conference Centre Shanghai, Shanghai, China Duur: 24 mei 2026 → 27 mei 2026 https://2026.ieee-iscas.org/

Conference

Conference	2026 IEEE International Symposium on Circuits and Systems
Verkorte titel	ISCAS 2026
Land/Regio	China
Stad	Shanghai
Periode	24/05/26 → 27/05/26
Internet adres	https://2026.ieee-iscas.org/

Duurzame ontwikkelingsdoelstellingen van de VN

Deze output draagt bij aan de volgende duurzame ontwikkelingsdoelstelling(en)

SDG 03 – Goede gezondheid en welzijn

Keywords

kenmerkselectie
gezondheidszorg
kleine datasets

Research Focus Areas Hanze University of Applied Sciences

Ondernemerschap
Energie
Healthy Ageing

Research Focus Areas Research Centre or Centre of Expertise

Digitale Transformatie

Publinova thema's

Taal, Cultuur & Kunsten
Economie en Management
Recht
ICT & Media
Gezondheid
Mens en Maatschappij
Techniek

Toegang tot document

10.1109/ISCAS66217.2026.11562723Licentie: Unspecified

Citeer dit

@conference{8dfbdf9dd64d47ca84763a38544705b4,

title = "Comparison of Feature Selection Methods for High-Dimensional Small Datasets",

abstract = "Machine learning (ML) techniques are widely applied in healthcare, ranging from precision healthcare to the possibility of preventive care. A general problem in clinical studies is the combination of high dimensionality and low sample sizes, complicating the process of extracting significant indicators from the feature set. In this paper, we compare the quality of three feature selection methods when the number of samples is low compared to the number of features. The results are compared to those of earlier research that used an optimization process applied to recursive feature elimination. The two filter methods, Anova and chi-squared, were unable to produce stable results, whereas the embedded Lasso method showed more promising outcomes.",

keywords = "feature selection, healthcare, small datasets, kenmerkselectie, gezondheidszorg, kleine datasets",

author = "Harald Rietdijk and Patricia Conde-Cespedes and Talko Dijkhuis and Hilbrand Oldenhuis and Maria Trocan",

year = "2026",

month = may,

day = "24",

doi = "10.1109/ISCAS66217.2026.11562723",

language = "English",

pages = "2893--2897",

note = "2026 IEEE International Symposium on Circuits and Systems ; Conference date: 24-05-2026 Through 27-05-2026",

url = "https://2026.ieee-iscas.org/",

}

TY - CONF

T1 - Comparison of Feature Selection Methods for High-Dimensional Small Datasets

AU - Rietdijk, Harald

AU - Conde-Cespedes, Patricia

AU - Dijkhuis, Talko

AU - Oldenhuis, Hilbrand

AU - Trocan, Maria

PY - 2026/5/24

Y1 - 2026/5/24

N2 - Machine learning (ML) techniques are widely applied in healthcare, ranging from precision healthcare to the possibility of preventive care. A general problem in clinical studies is the combination of high dimensionality and low sample sizes, complicating the process of extracting significant indicators from the feature set. In this paper, we compare the quality of three feature selection methods when the number of samples is low compared to the number of features. The results are compared to those of earlier research that used an optimization process applied to recursive feature elimination. The two filter methods, Anova and chi-squared, were unable to produce stable results, whereas the embedded Lasso method showed more promising outcomes.

AB - Machine learning (ML) techniques are widely applied in healthcare, ranging from precision healthcare to the possibility of preventive care. A general problem in clinical studies is the combination of high dimensionality and low sample sizes, complicating the process of extracting significant indicators from the feature set. In this paper, we compare the quality of three feature selection methods when the number of samples is low compared to the number of features. The results are compared to those of earlier research that used an optimization process applied to recursive feature elimination. The two filter methods, Anova and chi-squared, were unable to produce stable results, whereas the embedded Lasso method showed more promising outcomes.

KW - feature selection

KW - healthcare

KW - small datasets

KW - kenmerkselectie

KW - gezondheidszorg

KW - kleine datasets

U2 - 10.1109/ISCAS66217.2026.11562723

DO - 10.1109/ISCAS66217.2026.11562723

M3 - Paper

SP - 2893

EP - 2897

T2 - 2026 IEEE International Symposium on Circuits and Systems

Y2 - 24 May 2026 through 27 May 2026

ER -

Comparison of Feature Selection Methods for High-Dimensional Small Datasets

Samenvatting

Conference

Duurzame ontwikkelingsdoelstellingen van de VN

Keywords

Research Focus Areas Hanze University of Applied Sciences

Research Focus Areas Research Centre or Centre of Expertise

Publinova thema's

Toegang tot document

Vingerafdruk

Citeer dit