ABSTRACT
Medical expert systems will begin to appear as researchers in medical artificial intelligence continue to make progress in key problems such as general system architectures, management of uncertainty, knowledge acquisition and encoding, man-machine interface and system integration into clinical environments. An expert system architecture specializes common Al problem-solving techniques to a particular class of tasks. E. T. Keravnou and L. Johnson have argued that the more closely an expert system captures the underlying model of competence, the more satisfactory it will be as a component in a human-computer system. A data interpretation module transforms available patient data into the form used by the rest of the system for its reasoning. Normal ranges are determined from the statistics of a population, but the interpretation of a case requires one to establish whether a finding is abnormal in the individual.
