Lean Kernels in Description Logics

摘要: Lean kernels (LKs) are an effective optimization for deriving the causes of unsatisfiability a propositional formula. Interestingly, no analogous notion exists explaining consequences description logic (DL) ontologies. We introduce LKs DLs using general consequence-based methods, and provide algorithm computing them which incurs in only linear time overhead. As example, we instantiate our framework to DL \({\mathcal {ALC}}\). prove formally empirically that tighter approximation set relevant axioms consequence than syntactic locality-based modules.