A Refined Canonical Data Model for Multi-schema Integration and Mapping

Abstract

Current approaches in e-Business schema mapping and matching focus on pair-wise comparison of schemas. This paper gives an overview of how n-way comparison of e-Business schemas via a canonical data model can benefit to schema matching and mapping processes. We construct a refined canonical data model by integrating a set of input schemas into one comprehensive representation. This canonical data model represents the closure of all integrated schemas. In several use cases, the advantages of the canonical data model are demonstrated. Our work is based on excerpts from realistic input schemas and provides a concrete, ideal canonical data model as a reference for further research.