The Medical Diagnosis Based On Fuzzy Expert System Technology Research

This thesis applys fuzzy math theory,combines with some knowledge representation methods and reasoning techniques, and uses related technology of database to study key techonology of medical diagnosis expert system on basis of deficiency of current expert systems.The work of the paper focused on the following some aspects:(1) The paper perfects knowledge representation methods combining with production knowledge representation and frame knowledge representation,and get over the shortcoming of single knowledge representation method.(2) The paper applys forward reasoning and backward reasoning to complete fuzzy rule reasoning mechanism and frame reasoning mechanism base on process and characteristics of medical expert’s problem solving.(3) In order to convenience of reasoning, the paper divides attribute values into fuzzy continuous attribute and fuzzy discrete attribute on basis of characteristics of and fuzzy attributes. In reasoning process,the system applys MIN-MAX matching algorithm to calculate matching degree of two fuzzy continuous attributes and uses semantic distance to calculate matching degree of two fuzzy discrete attributes,which is more reasonable to calculate matching degree.(4) The paper applys breadth-first searching algorithm, and records the reasoning process and middle conclusion,while makes user get the explains according to his require.(5) The paper proposes the medical diagnosis expert system that is simple and useful,and while may be embed in other expert system.The system is tested and shows that knowledge representation method of medical diagnosis expert system in the paper is helpful for optimizing che knowledge base structure and achieving che relative independence of knowledge base maintance.It applys fuzzy rule reasoning mechanism and frame reasoning mechanism that may reason according to medical expert’s thingking process and characteristics of medical knowledge and improve the reasoning efficiency.