Anti-Unification of Concepts in Description Logic EL

We study anti-unification for the description logic EL and introduce the notion of least general generalisation, which generalises simultaneously least common subsumer and concept matching. The idea of generalisation of two concepts is to detect maximal similarities between them, and to abstract over their differences uniformly. We demonstrate that a finite minimal complete set of generalisations for EL concepts always exists and establish complexity bounds for computing them. We present an anti-unification algorithm that computes generalisations with a fixed skeleton, study its properties and report on preliminary experimental evaluation.