Lanzola G, Stefanelli M
Dipartimento di Informatica e Sistemistica, Università degli Studi di Pavia, Italy.
Comput Biomed Res. 1992 Aug;25(4):351-65. doi: 10.1016/0010-4809(92)90025-6.
For a knowledge-based system (KBS) to exhibit an intelligent behavior, it must be endowed with knowledge enabling it to represent the expert's strategies. The elicitation task is inherently difficult for strategic knowledge, because strategy is often tacit, and, even when it has been made explicit, it is not an easy task to describe it in a form which may be directly translated and implemented into a program. This paper describes a Specialized Framework for Medical Diagnostic Knowledge-Based Systems that can help an expert in the process of building KBSs in a medical domain. The framework is based on an epistemological model of diagnostic reasoning which has proven to be helpful in describing the diagnostic process in terms of the tasks that it is composed of. It allows a straightforward modeling of diagnostic reasoning at the knowledge level by the domain expert, thus helping to convey domain-dependent strategies into the target KBS.
对于一个基于知识的系统(KBS)而言,要展现出智能行为,它必须具备能够使其表示专家策略的知识。获取策略性知识的任务本身就很困难,因为策略通常是隐性的,而且即便将其明确表述出来,要以一种能够直接转化并实现为程序的形式来描述它也并非易事。本文描述了一种用于医学诊断基于知识系统的专门框架,该框架能够在医学领域帮助专家构建KBS的过程。此框架基于一种诊断推理的认识论模型,该模型已被证明有助于根据诊断过程所包含的任务来描述诊断过程。它允许领域专家在知识层面直接对诊断推理进行建模,从而有助于将领域相关策略转化到目标KBS中。