摘要
本文针对医疗诊断推理的难点,提出了以自反馈的AM网络为主、以规则推理为辅的面向对象的医疗诊断推理模型。这种模型既具有AM网络的优点,又克服了AM网络容量有限的缺陷。结合面向对象的知识封装,将学习机融于推理机,将推理机融于知识表示,在MDES的设计中,表现出良好的适应性。
According to the difficulties of medical - diagnostic inference, this paper uses OOP method to build a model which places auto-feedback AM network as a major inference mechanism and rules -based inference as an assistant. Having overcome the disadvantage in the little capacity of AMNN and utilizing the merits of AMNN and OOP, the model puts learning machine into inference machine and inference machine into knowledge representation. This model shows good adaptations to the designing of MDES.
出处
《计算机应用与软件》
CSCD
北大核心
2001年第1期5-9,42,共6页
Computer Applications and Software
