Compositional Instance-Based Learning

This paper proposes a new algorithm for acquisition of preference predicates by a learning apprentice, termed Compositional Instance-Based Learning (CIBL), that permits multiple instances of a preference predicate to be composed, directly exploiting the transitivity of preference predicates. In an empirical evaluation, CIBL was consistently more accurate than a I-NN instance-based learning strategy unable to compose instances. The relative performance of CIBL and decision tree induction was found to depend upon (1) the complexity of the preference predicate being acquired and (2) the dimensionality of the feature space.