Control issues in classificatory diagnosis
Jon Sticklen, Bharath Chandrasekaran, John R. Josephson · International Joint Conference on Artificial Intelligence · 1985
A good part of medical diagnosis can be modeled as classification problem solver producing a differential, working in conjunction with an abductive component that performs differential diagnosis by synthesising a best composite hypothesis out of the hypotheses in the differential list. Classification problem solving itself can be viewed as having a control component which selects hypotheses to consider, and a decision component associated with each selected hypothesis. In this paper we study the family of control regimes that are useful in classificatory problem solving. We start with MDX, a classification system organized as a hierarchical collection of hypothesis specialists, critique its control behavior, and by considering a set of situations involving multiple diseases, show how elements can be added to the control regime in a modular way to handle a large variety of situations.