Abstract: Viewing at the problems that the disease diagnosis mainly depends on the personal experience judgment and the rule-based reasoning inefficiency and lacking of flexibility, an algorithm of case matching is proposed. The diagnosis results are achieved through symptoms matching and analysis of patients, the recommended treatment is given depended on the results. The relations between the new symptoms and related disease are determined by the calculation of the correlation, and are updated into the case library. Obtained by experiment, the accuracy rate of system diagnosis is 90.18%, it shows that the system is of a high diagnostic efficiency and better intelligence, and can provide a certain supporting for doctors in medical diagnosis.


Key words: disease diagnosis, case-based reasoning, expert system, matching