A method for creating an enormous medical knowledge base.

Using a high-density knowledge representation method designed by us, we have developed the Enormous Knowledge Base of Disease Diagnosis Criteria (EKBDDC). It contains diagnostic criteria of 1001 diagnostic entities and describes nearly 4000 items of diagnostic indicators. It is the core of a huge medical project--Electronic Brain Medical Erudite (EBME). This enormous knowledge base was implemented initially on a low-cost popular microcomputer, which can aid in prompting of typical disease and in teaching of diagnosis. This knowledge base will be constantly expanded and adapted to the need of diagnosing of atypical diseases. By means of a software interface it will be connected with the international medical information systems. We have also explored an assembling technique of medical knowledge base. To test the behavior of EBME we performed a series of trials with a total of 815 cases. The diagnostic accordance rates were 89.7, 89.4, and 85%, respectively. It demonstrated that this system should be improved before clinical application.