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。 The article proposes a method of fuzzy mathematics and rule-based neural network to construct an expert system of respiratory diseases, including representation, acquisition and reasoning of knowledge. For fuzzy mathematical methods, fuzzy sets are used to represent the symptoms considered and all possible illnesses. Medical knowledge is stored in the fuzzy relationship between symptoms and illness. The method of using fuzzy relationship synthesis when inference. For the rule-based neural network method, the structure of the network is automatically constructed from the rule set to determine the number of hidden nodes and the connection weights. Using parallel methods to reason.