An Empirical Comparison of Three Inference Methods
This work addresses the selection of inference methods for medical diagnosis systems, but it is incremental as it focuses on empirical comparison without introducing new techniques.
The paper tackled the problem of comparing three inference methods for uncertain reasoning in the Pathfinder expert system for lymph-node pathology diagnosis, finding that the methods were evaluated using expert-rating and decision-theoretic metrics to assess diagnostic accuracy.
In this paper, an empirical evaluation of three inference methods for uncertain reasoning is presented in the context of Pathfinder, a large expert system for the diagnosis of lymph-node pathology. The inference procedures evaluated are (1) Bayes' theorem, assuming evidence is conditionally independent given each hypothesis; (2) odds-likelihood updating, assuming evidence is conditionally independent given each hypothesis and given the negation of each hypothesis; and (3) a inference method related to the Dempster-Shafer theory of belief. Both expert-rating and decision-theoretic metrics are used to compare the diagnostic accuracy of the inference methods.