Аннотация:Compositional inference is an approach for approximate reasoning. This kind of inference is applied in the building of rule-based compositional systems for medicine such as MYCIN and CADIAG-2. These systems are rule based, that is, their knowledge consists of rules of the form IF (premise) THEN (conclusion) with some weight (degree of belief), and compositional, that is, they combine effects of particular rules using a binary combining function to compute their joint effect. The paper proposes an approach to applying approximate reasoning for oriental traditional medical rule-based compositional systems. In general, the diagnosis of oriental traditional medicine includes a syndrome differentiation (BatCuong), and internal organ (TangPhu) diagnosis etc. The diagnosis process in oriental traditional medicine is called BienChung. As the pathogenesis (BatCuong, TangPhu, etc.) is convinced, then the prescription can be given. The result of BienChung may be the composition of two or more syndromes of pathogenesis. The author describes the diagnosis process of oriental traditional medicine (OTM) using compositions of rules as an inference mechanism model of the BienChung system of OTM. It accepts fuzzy descriptions of the patient's symptoms and their fuzzy relationships in the form of rules and the negation of rules then infers descriptions of diagnoses as BatCuong, TangPhu, pathogenesis composition, BatCuong and TangPhu etc. according to the theory of oriental traditional medicine.
