UNIFYING INSTANCE-BASED AND RULE-BASED INDUCTION

Several well-developed approaches to inductive learning now exist, but each has specific limitations that are hard to overcome. Multi-strate gy learning attempts to tackle this problem by combining multiple meth ods in one algorithm. This article describes a unification of two wide ly-used empirical approaches: rule induction and instance-based learni ng. In the new algorithm, instances are treated as maximally specific rules, and classification is performed using a best-match strategy. Ru les are learned by gradually generalizing instances until no improveme nt in apparent accuracy is obtained. Theoretical analysis shows this a pproach to be efficient. It is implemented in the RISE 3.1 system. In an extensive empirical study, RISE consistently achieves higher accura cies than state-of-the-art representatives of both its parent approach es (PEBLS and CN2), as well as a decision tree learner (C4.5). Lesion studies show that each of RISE's components is essential to this perfo rmance. Most significantly, in 14 of the 30 domains studied, RISE is m ore accurate than the best of PEELS and CN2, showing that a significan t synergy can be obtained by combining multiple empirical methods.