Medical Knowledge-Based System for Diagnosis from Symptoms and Signs

This research proposes design and development of a medical knowledge-based system (MKBS) for diagnosis from symptoms and signs. This system was developed to support a knowledge construction and an inference engine. The knowledge construction was based on a concept of production rules, which was performed in tree structure. The inference engine used interactive forward chaining technique to infer a diagnostic result. The proposed system was designed to interact with user by using question forms of symptoms, and it was able to support text and picture information. The architecture of this system consisted of inference engine, knowledge base, user interface unit, knowledge acquisition unit, explanation module and blackboard. This medical knowledge-based system was developed by Borland c language on Windows XP system. Finally this system was applied to check accuracy by comparing with general physician’s diagnoses. The experimental result showed the diagnosis of the system more than 97% accurately at the 0.01 level of significance, when it is compared with diagnosis of a physician.