EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Modeling health risks using neural network ensembles

Brandon M Smith iD, Antonio Criminisi, Noam Sorek, Yaar Harari, Neeraj Sood, Steven B Heymsfield

PLOS ONE · 09.10.2024 · 19 (10) · e0308922

DOI: 10.1371/journal.pone.0308922 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS ONE, online erschienen am 09.10.2024. Verfasst von Brandon M Smith und 5 weiteren Autor:innen. Beteiligte Einrichtung: Amazon.com, LLC, Washington, D. C, United States of America. Hauptthemen laut MEDLINE: Neural Networks, Computer, Body Mass Index, Obesity. Die Arbeit wurde bisher 7-mal zitiert (Crossref) und nennt 53 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: NIDDK NIH HHS, Funded by Amazon, Inc., Seattle, WA..

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This study aims to demonstrate that demographics combined with biometrics can be used to predict obesity related chronic disease risk and produce a health risk score that outperforms body mass index (BMI)—the most commonly used biomarker for obesity. We propose training an ensemble of small neural networks to fuse demographics and biometrics inputs. The categorical outputs of the networks are then turned into a multi-dimensional risk map, which associates diverse inputs with stratified, output health risk. Our ensemble model is optimized and validated on disjoint subsets of nationally representative data (N~100,000) from the National Health and Nutrition Examination Survey (NHANES). To broaden applicability of the proposed method, we consider only non-invasive inputs that can be easily measured through modern devices. Our results show that: (a) neural networks can predict individual conditions (e.g., diabetes, hypertension) or the union of multiple (e.g., nine) health conditions; (b) Softmax model outputs can be used to stratify individual- or any-condition risk; (c) ensembles of neural networks improve generalizability; (d) multiple-input models outperform BMI (e.g., 75.1% area under the receiver operator curve for eight-input, any-condition models compared to 64.2% for BMI); (e) small neural networks are as effective as larger ones for the inference tasks considered; the proposed models are small enough that they can be expressed as human-readable equations, and they can be adapted to clinical settings to identify high-risk, undiagnosed populations.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Neural Networks, ComputerBody Mass IndexObesityHumansNutrition SurveysFemaleMaleRisk FactorsRisk AssessmentAdultMiddle Aged

Autor:innen und Institutionen

  • Amazon.com, LLC, Washington, D. C, United States of America.
  • Amazon UK Services Ltd., London, United Kingdom.
  • Amazon.com, LLC, Israel.
  • USC Sol Price School of Public Policy, Los Angeles, CA, United States of America.
  • Pennington Biomedical Research Center, Louisiana State University System, Baton Rouge, Louisiana, United States of America.

ORCID-Profile: Brandon M Smith

Förderung

  • NIDDK NIH HHS
  • Funded by Amazon, Inc., Seattle, WA.

Zitieren

APA

Smith, B. M., Criminisi, A., Sorek, N., Harari, Y., Sood, N., & Heymsfield, S. B. (2024). Modeling health risks using neural network ensembles. PLOS ONE, 19(10), e0308922. https://doi.org/10.1371/journal.pone.0308922

Vancouver

Smith BM, Criminisi A, Sorek N, Harari Y, Sood N, Heymsfield SB. Modeling health risks using neural network ensembles. PLOS ONE. 2024;19(10):e0308922. doi:10.1371/journal.pone.0308922

BibTeX
@article{Smith2024,
  author = {Smith, Brandon M and Criminisi, Antonio and Sorek, Noam and Harari, Yaar and Sood, Neeraj and Heymsfield, Steven B},
  title = {Modeling health risks using neural network ensembles},
  journal = {PLOS ONE},
  year = {2024},
  volume = {19},
  number = {10},
  pages = {e0308922},
  doi = {10.1371/journal.pone.0308922}
}
Kostenlos starten →