Predicate calculus, artificial intelligence, and workers' compensation.
J Occup Med 1989 May; 31(5):484-489
An expert system to assist practitioners in recognizing occupational pulmonary interstitial disease was presented as a model to evaluate the utility of predicate calculus, fuzzy set theory, and artificial intelligence in clinical decision making. The sources of clinical uncertainty include competing causation, or attributable risk; the uncertainty inherent in a clause or group of clauses of a predicate; the degree of certainty of the entire rule, rather than of a particular clause; and the need to combine certainties to obtain an overall confidence factor. Uncertainty can be addressed by predicate calculus, fuzzy set theory, and the search procedures used in artificial intelligence. One such procedure, backward chaining, requires the specification of a goal, which is validated by the knowledge base and rule representation of the expert system. The other procedure, forward chaining, is driven by data input to the system, inferring from that data an appropriate set of implications. Although both are valid, these inferential procedures are not equivalent, and may lead to different conclusions. The authors conclude that predicate calculus, fuzzy set theory, and artificial intelligence techniques are useful in assisting with clinical diagnosis, in creating a sense of the diagnostic process, and in resolving conflicts between competent diagnosticians.
JOCMA7; NIOSH-Publication; NIOSH-Grant; Analytical-models; Clinical-diagnosis; Computer-models; Information-processing; Lung-fibrosis; Occupational-respiratory-disease; Pulmonary-system-disorders
Medicine University of California 405 Hilgard Avenue Los Angeles, CA 90024
Journal of Occupational Medicine
University of California Los Angeles, Los Angeles, California