Machine Learning for Data Mining in Medicine

  • Authors:
  • Nada Lavrac

  • Affiliations:
  • -

  • Venue:
  • AIMDM '99 Proceedings of the Joint European Conference on Artificial Intelligence in Medicine and Medical Decision Making
  • Year:
  • 1999

Quantified Score

Hi-index 0.00

Visualization

Abstract

Large collections of medical data are a valuable resource from which potentially new and useful knowledge can be discovered through data mining. This paper gives an overview of machine learning approaches used in mining of medical data, distinguishing between symbolic and sub-symbolic data mining methods, and giving references to applications of these methods in medicine. In addition, the paper presents selected measures for performance evaluation used in medical prediction and classification problems, proposing also some alternative measures for rule evaluation that can be used in ranking and filtering of induced rule sets.