Published: 2012 Mai
Type: Research Notes
Institution: AIFB, KIT
Erscheinungsort / Ort: Karlsruhe
In this paper, we consider the task of knowledge base extraction with its three conflicting requirements: the size of the extracted knowledge base, the size of the corresponding signature and the syntactic similarity of the extracted knowledge base with the originally given one. We demonstrate that, both, minimal module extraction and uniform interpolation, assign an absolute priority to one of these requirements, thereby limiting the possibilities to influence the other two. To account for scenarios, in which such an extreme prioritization is not necessary, we investigate the task of knowledge base extraction for EL based on two alternative, less restrictive notions of syntactic similarity with the second highest priority given to the knowledge base size. Moreover, to address scenarios, where computation time is important, we propose a tractable rewriting approach based on the chosen prioritization of requirements and empirically compare this novel technique with the existing implemented approaches with encouraging results.