Generalization challenges in electrocardiogram deep learning: insights from dataset characteristics and attention mechanism.

Journal: Future cardiology
Published Date:

Abstract

Deep learning's widespread use prompts heightened scrutiny, particularly in the biomedical fields, with a specific focus on model generalizability. This study delves into the influence of training data characteristics on the generalization performance of models, specifically in cardiac abnormality detection. Leveraging diverse electrocardiogram datasets, models are trained on subsets with varying characteristics and subsequently compared for performance. Additionally, the introduction of the attention mechanism aims to improve generalizability. Experiments reveal that using a balanced dataset, just 1% of a large dataset, leads to equal performance in generalization tasks, notably in detecting cardiology abnormalities. This balanced training data notably enhances model generalizability, while the integration of the attention mechanism further refines the model's ability to generalize effectively.

Authors

  • Zhaojing Huang
    College of Chemistry, Chemical Engineering & Environmental Science, Minnan Normal University, Zhangzhou 363000, China.
  • Sarisha MacLachlan
    School of Biomedical Engineering at The University of Sydney, Camperdown, NSW 2006, Australia.
  • Leping Yu
    School of Biomedical Engineering at The University of Sydney, Camperdown, NSW 2006, Australia.
  • Luis Fernando Herbozo Contreras
    School of Biomedical Engineering at The University of Sydney, Camperdown, NSW 2006, Australia.
  • Nhan Duy Truong
    School of Electrical and Information Engineering, University of Sydney, Sydney, NSW 2006, Australia; Nano-Neuro-inspired Research Laboratory, School of Electrical and Information Engineering, University of Sydney, Sydney, NSW 2006, Australia; School of Engineering, Royal Melbourne Institute of Technology, Melbourne, VIC 3000, Australia. Electronic address: duy.truong@sydney.edu.au.
  • Antonio Horta Ribeiro
    Department of Information Technology at Uppsala University, 753 10 Uppsala, Sweden.
  • Omid Kavehei
    School of Electrical and Information Engineering, University of Sydney, Sydney, NSW 2006, Australia; Nano-Neuro-inspired Research Laboratory, School of Electrical and Information Engineering, University of Sydney, Sydney, NSW 2006, Australia. Electronic address: omid.kavehei@sydney.edu.au.