Predictive Abilities of Machine Learning Techniques May Be Limited by Dataset Characteristics: Insights From the UNOS Database.

Journal: Journal of cardiac failure
Published Date:

Abstract

BACKGROUND: Traditional statistical approaches to prediction of outcomes have drawbacks when applied to large clinical databases. It is hypothesized that machine learning methodologies might overcome these limitations by considering higher-dimensional and nonlinear relationships among patient variables.

Authors

  • P Elliott Miller
    Section of Cardiovascular Medicine, Yale School of Medicine, New Haven, Connecticut.
  • Sumeet Pawar
    Section of Cardiovascular Medicine, Yale School of Medicine, New Haven, Connecticut.
  • Benjamin Vaccaro
    Section of Cardiovascular Medicine and Center for Outcomes Research, Yale University School of Medicine New Haven, CT.
  • Megan McCullough
    Section of Cardiovascular Medicine, Yale School of Medicine, New Haven, Connecticut.
  • Pooja Rao
    Qure.ai, Mumbai, India.
  • Rohit Ghosh
    Qure.ai, Mumbai, India.
  • Prashant Warier
    Qure.ai, Mumbai, India.
  • Nihar R Desai
    Section of Cardiovascular Medicine and Center for Outcomes Research, Yale University School of Medicine New Haven, CT.
  • Tariq Ahmad
    Section of Cardiovascular Medicine and Center for Outcomes Research, Yale University School of Medicine New Haven, CT tariq.ahmad@yale.edu.