Instance Selection Algorithms for Predictive Modelling in Telehealth Applications

Fabian Wiesmüller; Dieter Hayn; Florian Hoffmann; Sten Hanke; Peter Kastner; Markus Falgenhauer; Günter Schreier

doi:10.3233/SHTI231083

Instance Selection Algorithms for Predictive Modelling in Telehealth Applications

Fabian Wiesmüller, Dieter Hayn, Florian Hoffmann, Sten Hanke, Peter Kastner, Markus Falgenhauer, Günter Schreier

Digital Health Information Systems

Research output: Chapter in Book or Conference Proceedings › Conference Proceedings with Oral Presentation › peer-review

Abstract

Telehealth services are becoming more and more popular, leading to an increasing amount of data to be monitored by health professionals. Machine learning can support them in managing these data. Therefore, the right machine learning algorithms need to be applied to the right data. We have implemented and validated different algorithms for selecting optimal time instances from time series data derived from a diabetes telehealth service. Intrinsic, supervised, and unsupervised instance selection algorithms were analysed. Instance selection had a huge impact on the accuracy of our random forest model for dropout prediction. The best results were achieved with a One Class Support Vector Machine, which improved the area under the receiver operating curve of the original algorithm from 69.91 to 75.88 %. We conclude that, although hardly mentioned in telehealth literature so far, instance selection has the potential to significantly improve the accuracy of machine learning algorithms.

Original language	English
Title of host publication	MedInfo 2023: The 19th World Congress on Medical and Health Informatics
Editors	Jen Bichel-Findlay, Paula Otero, Philip Scott, Elaine Huesing
Publisher	IOS Press BV
Pages	840-844
Number of pages	5
Volume	310
ISBN (Electronic)	978-1-64368-457-4
ISBN (Print)	978-1-64368-456-7
DOIs	https://doi.org/10.3233/SHTI231083
Publication status	Published - 25 Jan 2024
Event	MedInfo 2023: THE FUTURE IS ACCESSIBLE - International Convention Centre (ICC), Sydney, Australia Duration: 8 Jul 2023 → 12 Jul 2023 Conference number: 19. https://medinfo2023.org/

Conference

Conference	MedInfo 2023: THE FUTURE IS ACCESSIBLE
Abbreviated title	MedInofo 2023
Country/Territory	Australia
City	Sydney
Period	8/07/23 → 12/07/23
Internet address	https://medinfo2023.org/

Research Field

Exploration of Digital Health

Keywords

Instance selection
training data selection
predictive modelling
telehealth

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.3233/SHTI231083Licence: CC BY-NC

Cite this

Wiesmüller, F., Hayn, D., Hoffmann, F., Hanke, S., Kastner, P., Falgenhauer, M., & Schreier, G. (2024). Instance Selection Algorithms for Predictive Modelling in Telehealth Applications. In J. Bichel-Findlay, P. Otero, P. Scott, & E. Huesing (Eds.), MedInfo 2023: The 19th World Congress on Medical and Health Informatics (Vol. 310, pp. 840-844). IOS Press BV. https://doi.org/10.3233/SHTI231083

@inproceedings{9302256be1d5464592d04d00b28e1828,

title = "Instance Selection Algorithms for Predictive Modelling in Telehealth Applications",

abstract = "Telehealth services are becoming more and more popular, leading to an increasing amount of data to be monitored by health professionals. Machine learning can support them in managing these data. Therefore, the right machine learning algorithms need to be applied to the right data. We have implemented and validated different algorithms for selecting optimal time instances from time series data derived from a diabetes telehealth service. Intrinsic, supervised, and unsupervised instance selection algorithms were analysed. Instance selection had a huge impact on the accuracy of our random forest model for dropout prediction. The best results were achieved with a One Class Support Vector Machine, which improved the area under the receiver operating curve of the original algorithm from 69.91 to 75.88 %. We conclude that, although hardly mentioned in telehealth literature so far, instance selection has the potential to significantly improve the accuracy of machine learning algorithms.",

keywords = "Instance selection, training data selection, predictive modelling, telehealth",

author = "Fabian Wiesm{\"u}ller and Dieter Hayn and Florian Hoffmann and Sten Hanke and Peter Kastner and Markus Falgenhauer and G{\"u}nter Schreier",

note = "{\textcopyright} 2024 International Medical Informatics Association (IMIA) and IOS Press. This article is published online with Open Access by IOS Press and distributed under the terms of the Creative Commons Attribution Non-Commercial License 4.0 (CC BY-NC 4.0).; MedInfo 2023: THE FUTURE IS ACCESSIBLE, MedInofo 2023 ; Conference date: 08-07-2023 Through 12-07-2023",

year = "2024",

month = jan,

day = "25",

doi = "10.3233/SHTI231083",

language = "English",

isbn = " 978-1-64368-456-7",

volume = "310",

pages = "840--844",

editor = "Jen Bichel-Findlay and Paula Otero and Philip Scott and Elaine Huesing",

booktitle = "MedInfo 2023: The 19th World Congress on Medical and Health Informatics",

publisher = "IOS Press BV",

address = "Netherlands",

url = "https://medinfo2023.org/",

}

Wiesmüller, F , Hayn, D, Hoffmann, F, Hanke, S, Kastner, P, Falgenhauer, M & Schreier, G 2024, Instance Selection Algorithms for Predictive Modelling in Telehealth Applications. in J Bichel-Findlay, P Otero, P Scott & E Huesing (eds), MedInfo 2023: The 19th World Congress on Medical and Health Informatics. vol. 310, IOS Press BV, pp. 840-844, MedInfo 2023: THE FUTURE IS ACCESSIBLE, Sydney, Australia, 8/07/23. https://doi.org/10.3233/SHTI231083

Instance Selection Algorithms for Predictive Modelling in Telehealth Applications. / Wiesmüller, Fabian ; Hayn, Dieter; Hoffmann, Florian et al.
MedInfo 2023: The 19th World Congress on Medical and Health Informatics. ed. / Jen Bichel-Findlay; Paula Otero; Philip Scott; Elaine Huesing. Vol. 310 IOS Press BV, 2024. p. 840-844.

Research output: Chapter in Book or Conference Proceedings › Conference Proceedings with Oral Presentation › peer-review

TY - GEN

T1 - Instance Selection Algorithms for Predictive Modelling in Telehealth Applications

AU - Wiesmüller, Fabian

AU - Hayn, Dieter

AU - Hoffmann, Florian

AU - Hanke, Sten

AU - Kastner, Peter

AU - Falgenhauer, Markus

AU - Schreier, Günter

N1 - Conference code: 19.

PY - 2024/1/25

Y1 - 2024/1/25

N2 - Telehealth services are becoming more and more popular, leading to an increasing amount of data to be monitored by health professionals. Machine learning can support them in managing these data. Therefore, the right machine learning algorithms need to be applied to the right data. We have implemented and validated different algorithms for selecting optimal time instances from time series data derived from a diabetes telehealth service. Intrinsic, supervised, and unsupervised instance selection algorithms were analysed. Instance selection had a huge impact on the accuracy of our random forest model for dropout prediction. The best results were achieved with a One Class Support Vector Machine, which improved the area under the receiver operating curve of the original algorithm from 69.91 to 75.88 %. We conclude that, although hardly mentioned in telehealth literature so far, instance selection has the potential to significantly improve the accuracy of machine learning algorithms.

AB - Telehealth services are becoming more and more popular, leading to an increasing amount of data to be monitored by health professionals. Machine learning can support them in managing these data. Therefore, the right machine learning algorithms need to be applied to the right data. We have implemented and validated different algorithms for selecting optimal time instances from time series data derived from a diabetes telehealth service. Intrinsic, supervised, and unsupervised instance selection algorithms were analysed. Instance selection had a huge impact on the accuracy of our random forest model for dropout prediction. The best results were achieved with a One Class Support Vector Machine, which improved the area under the receiver operating curve of the original algorithm from 69.91 to 75.88 %. We conclude that, although hardly mentioned in telehealth literature so far, instance selection has the potential to significantly improve the accuracy of machine learning algorithms.

KW - Instance selection

KW - training data selection

KW - predictive modelling

KW - telehealth

U2 - 10.3233/SHTI231083

DO - 10.3233/SHTI231083

M3 - Conference Proceedings with Oral Presentation

C2 - 38269927

SN - 978-1-64368-456-7

VL - 310

SP - 840

EP - 844

BT - MedInfo 2023: The 19th World Congress on Medical and Health Informatics

A2 - Bichel-Findlay, Jen

A2 - Otero, Paula

A2 - Scott, Philip

A2 - Huesing, Elaine

PB - IOS Press BV

T2 - MedInfo 2023: THE FUTURE IS ACCESSIBLE

Y2 - 8 July 2023 through 12 July 2023

ER -

Instance Selection Algorithms for Predictive Modelling in Telehealth Applications

Abstract

Conference

Research Field

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this