IMPROVING ACCURACY BY COMBINING RULE-BASED AND CASE-BASED REASONING

An architecture is presented for combining rule-based and case-based r easoning. The architecture is intended for domains that are understood reasonably well, but still imperfectly. It uses a set of rules, which are taken to be only approximately correct, to obtain a preliminary a nswer for a given problem; it then draws analogies from cases to handl e exceptions to the rules, Having rules together with cases not only i ncreases the architecture's domain coverage, it also allows innovative ways of doing case-based reasoning: the same rules that are used for rule-based reasoning are also used by the case-based component to do c ase indexing and case adaptation. The architecture was applied to the task of name pronunciation, and, with minimal knowledge engineering, w as found to perform almost at the level of the best commercial systems . Moreover, its accuracy was found to exceed what it could have achiev ed with rules or cases alone, thus demonstrating the accuracy improvem ent afforded by combining rule-based and case-based reasoning.