학술논문

A methodology for validating intelligent systems in critical domains
Document Type
Conference
Source
Proceedings of 18th Annual International Conference of the IEEE Engineering in Medicine and Biology Society Engineering in medicine and biology Engineering in Medicine and Biology Society, 1996. Bridging Disciplines for Biomedicine. Proceedings of the 18th Annual International Conference of the IEEE. 5:2026-2027 vol.5 1996
Subject
Bioengineering
Intelligent systems
Performance analysis
Artificial intelligence
Application software
Knowledge representation
Sampling methods
Computer science
Computerized monitoring
Patient monitoring
Ventilation
Language
Abstract
In this work the authors present a formal approach to carrying out an analysis of the performance of intelligent monitoring systems (IMS) in intensive care units (ICUs). In this approach the use of any of the existing paradigms depends on the particular characteristics of the domain, and on the architecture of the IMS. The proposed methodology is based on previous retrospective analysis in which quantitative techniques of validation are applied to investigate the performance of the system. This methodology has been successfully applied to evaluate PATRICIA, an intelligent monitoring system designed to guide clinicians in handling ICU patients long-term dependent on mechanical ventilation.