EvoPat – Pattern-Based Evolution and Refactoring of RDF Knowledge Bases

Abstract. Facilitating the seamless evolution of RDF knowledge bases on the Semantic Web presents still a major challenge. In this work we devise EvoPat – a pattern-based approach for the evolution and refactor-ing of knowledge bases. The approach is based on the definition of basic evolution patterns, which are represented declaratively and can capture simple evolution and refactoring operations on both data and schema lev-els. For more advanced and domain-specific evolution and refactorings, several simple evolution patterns can be combined into a compound one. We performed a comprehensive survey of possible evolution patterns with a combinatorial analysis of all possible before/after combinations, result-ing in an extensive catalog of usable evolution patterns. Our approach was implemented as an extension for the OntoWiki semantic collabora-tion platform and framework.