Chin H L, Cooper G F
Permanente Medical Group, Inc., San Jose, CA 95119-1197.
Comput Methods Programs Biomed. 1989 Oct-Nov;30(2-3):185-98. doi: 10.1016/0169-2607(89)90071-0.
The past decade has seen the emergence of programs that make use of large knowledge bases to assist physicians in diagnosis within the general field of internal medicine. One such program, Internist-I, contains knowledge about over 600 diseases, covering a significant proportion of internal medicine. This paper describes the process of converting a subset of this knowledge base--in the area of cardiovascular diseases--into a probabilistic format, and the use of this resulting knowledge base to teach medical diagnostic knowledge. The system (called KBSimulator--for Knowledge-Based patient Simulator) generates simulated patient cases and uses these cases as a focal point from which to teach medical knowledge. This project demonstrates the feasibility of building an intelligent, flexible instructional system that uses a knowledge base constructed primarily for medical diagnosis.
在过去十年中，出现了一些利用大型知识库协助内科医生进行诊断的程序。其中一个程序，即内科医生一号（Internist-I），包含了600多种疾病的知识，涵盖了内科的很大一部分。本文描述了将该知识库中心血管疾病领域的一个子集转换为概率格式的过程，以及使用由此产生的知识库来传授医学诊断知识的情况。该系统（称为基于知识的患者模拟器KBSimulator）生成模拟患者病例，并将这些病例作为传授医学知识的重点。该项目证明了构建一个智能、灵活的教学系统的可行性，该系统使用主要为医学诊断构建的知识库。
