Incremental Learning of Rules and Meta-rules

Marc Schoenauer, Michèle Sebag

Abstract: This paper presents a general incremental learning scheme: a single generalization algorithm can both earn a set of rules from a set of examples, and achieve the refinement of a previous set of rules. This approach is based on a redescription operator called reduction: from a set of examples and a set of rules, we derive a new set of examples describing the behavior of the rule set. New rules are extracted from these behavioral examples: those rules can be seen as meta-rules, as they control previous rules in order to improve their predictive accuracy.