Problems and solutions in the development of an expert system

Three notable sources of fuzziness in the laboratory field are discussed. These are laboratory data themselves, the degree to which test parameters participate in the diagnosis, and the expression of diagnostic reliability. To solve these problems of fuzziness, artificial intelligence rules have been used. Verification of the knowledge base was performed using real laboratory data from 155 patients who were diagnosed as having hematological disorders by expert hematologists. After the diagnostic process was executed using the expert system, the responses were compared with the actual diagnoses. The results indicated useful improvements in hematological data analysis