医学专家系统中知识表示、获取和推理的两种方法

Two Methods of Knowledge Representation,Knowledge Acquisition and Knowledge Reasoning for a Medical Expert System

文章提出使用模糊数学的方法和基于规则的神经网络的方法来构造一个呼吸道疾病方面的专家系统,包括知识的表示、获取和推理。对模糊数学方法,用模糊集来表示所考虑的症状与所有可能的疾病。医学知识存储在症状与疾病的模糊关系上。推理时使用模糊关系合成的方法。对基于规则的神经网络方法,从规则集中自动构造网络的结构,确定隐层节点数和连接权值。用并行的方法进行推理。

This paper presents two methods of knowledge representation,knowledge acquisition and knowledge reason-ing for a expert system of the diagnoses of respiratory disease.One is to represent the disease condition and the all correlative disease with fuzzy sets.The clinic knowledge is stored in the fuzzy relations between disease condition and the all correlative disease,and It can get the clinical diagnosis from these fuzzy relations by synthesizing.Another method is to use the rule-based neural network to construct the neural network's structure and its weights by collect-ing the rules automatically,and to do the knowledge reasoning with parallel method.