An approach for the incorporation of high level knowledge into clinical alarm systems.

Title: An approach for the incorporation of high level knowledge into clinical alarm systems.
Identifier: AAI9009740
identifier: 9009740
Creator: Henkind, Steven Joseph.
Contributor: Adviser: Louis E. Teichholz
Date: 1989
Language: English
Publisher: City University of New York.
Subject: Computer Science | Health Sciences, Medicine and Surgery
Abstract: Clinical alarm systems have been well accepted in medical practice but suffer from several major limitations; for example, these systems are generally not able to handle patient/disease specificity, temporal changes, multivariable combinations, and dynamic patterns. In order for an alarm system to overcome these limitations, it is necessary that the system utilize high level knowledge.;In this thesis we develop an approach for incorporating such knowledge into alarm systems. A fundamental aspect of our approach is the use of characteristic functions. In particular, we demonstrate that characteristic functions (a notion from fuzzy set theory) can serve as a powerful knowledge representation for knowledge based systems, even if fuzzy set operators are not utilized.;A prototype alarm system was developed using our approach; the system is specifically oriented to the detection of cardiac tamponade in cardiac surgical intensive care unit patients. This system exhibited promising behavior in limited tests.
Type: dissertation
Source: PQT Legacy CUNY.xlsx
degree: Ph.D.

Media: An approach for the incorporation of high level knowledge into clinical alarm systems.