Qualitative modeling as a paradigm for diagnosis and prediction in critical care environments

  • Authors:

  • Serdar Uckun;Benoit M. Dawant

  • Affiliations:

  • Department of Biomedical Engineering, Vanderbilt University, Box 1804 Station B, Nashville, TN 37235, USA;Department of Electrical Engineering, Vanderbilt University, Box 1824 Station B, Nashville, TN 37235, USA

  • Venue:
  • Artificial Intelligence in Medicine
  • Year:
  • 1992

Quantified Score

Hi-index 0.00

Visualization

Abstract

This article describes ongoing research for the development of a physiological model for diagnosis, prediction, and therapy management purposes. The patient model is part of an intelligent monitoring system designed for intensive care applications, and it is based on ontology which extends the Qualitative Process theory. A major feature of the modeling ontology is the capability to incorporate numerical values in qualitative simulations in addition to purely qualitative values. The article also discusses the use of model-based reasoning for diagnosis and prediction tasks in physiological domains. Preliminary efforts for validation of the model are described. The current limitations of the patient model and the framework are outlined, and future research directions are stated.