Learning and classifying

Abstract

We define and study a learning paradigm that sits between identification in the limit and classification. More precisely, we expect that a learner be able to identify in the limit which members of a set D of n possible data belong to a target language, where n and D are arbitrary. We show that Ex- and BC-learning are often more difficult than performing this classification task, taking into account desirable constraints on how the learner behaves, such as bounding the number of mind changes and being conservative. Special attention is given to various forms of consistency. We provide a fairly comprehensive set of results that demonstrate the fruitfulness of the approach and the richness of the paradigm. © 2011 Springer-Verlag.