The Study On Clinical Knowledge Representation And Reasoning Mechanism Of Osteoma Computer-aided Diagnostic Expert System

Osteoma computer-aided diagnostic expert system is a series of programs based on theories and methods of expert system, and simulates the progresses of medical experts' thinking of diagnosing and treating osteoma. As an assistant implement, it has some fine characteristics, such as illumination, transparence and flexibility. It can continue working exactly and quickly not affected by time, space and conditions and can improve the efficiency. The work efficiency of the system is high. So it can help osteal doctors solve complicated diagnostic problems and can enhance the diagnostic accuracy.Knowledge base and inference engine are the two main parts of expert system. The technology of knowledge representation is importantand is one of the most fundamental questions in the knowledge engineering. Also it is an important topic of the study on expert system. The mode selection of knowledge representation correlates knowledge storage, an ability of acquiring knowledge and the efficiency of using knowledge.Aiming at the characteristics of osteoma disease, first, we understood the diagnostic manner of osteoma in this research. By discussing and communing with the domain experts, we obtained the cases and the believable data that are needed by clinical diagnosis; secondly, according as the osteoma knowledge and the characteristics of the data, we used a technology of hybrid knowledge representation to express the diagnostic information on osteoma. Based on the known methods of knowledge representation, we improved the method of our system and proposed a method of combining frame knowledge representation with heuristic rule knowledge representation. At the same time, a feasible method was put forward, combining database with logical programming, to express knowledge by computer programming. And then an antetype of the knowledge base was built. Finally, a corresponding inference engine was studied. The reasoning process was implemented with using the internal reasoning mechanism in Prolog and fuzzy method. The system was tested by several kinds of osteoma cases. The result showed that the knowledge base could store diagnostic information effectively and represent knowledge correctly. The operation of computer is simple so that it could save time. Moreover, our system can do its work. But the reasoning mechanism needs to be improved continually with the accession of the cases and farther classification of the osteoma. Likedoing that, the diagnostic veracity will be heightened and the diagnostic level will be enhanced.The main new ideas in this paper are as follows: 1. Based on theories of expert system and knowledge engineering, a hybrid knowledge representation that was accorded with clinical diagnostic knowledge on osteoma was studied and put forward; 2. A method was proposed combining database with programming in logic to build an antetype of the knowledge base of our system. We provided an effective method for building a similar system; 3. The reasoning mechanism with fuzzy theory was studied and an inference engine was implemented successfully that could simulate the process of diagnosis and accorded with clinical knowledge representation of osteoma.