Ontology evolution in data integration

Abstract

Due to the rapid scientific development, ontologies and schemata need to change. When ontologies evolve, the changes should somehow be rendered and used by the pre-existing data integration systems. In most of these systems, when ontologies change their relations with the data sources i.e. the mappings, are recreated manually, a process which is known to be error-prone and time-consuming.In this dissertation, we provide a solution that allows query answering under evolving ontologies without mapping redefinition. This is achieved by exploiting query rewriting. We elegantly separate the semantics of query rewriting for different ontology versions and for the sources and we present a module that enables ontology evolution over traditional ontology-based data integration systems. That module gets as input the different ontology versions and the user query, and rewrites the query over data integration systems that use different ontology versions. This is performed by the automatic identification of changes among the ontology versions using a high-level language of changes, which are then interpreted as GAV mappings to enable query rewriting among ontology versions.Although query rewriting always succeeds, several problems may occur due to non information preserving changes among the ontology versions. We identify the problems in such a setting and we provide efficient, intuitive solutions, either by explaining the reasons for the failure or by producing the best “over-approximations”.We prove that our approach imposes only a small overhead over traditional query rewriting algorithms and it is modular and scalable. Finally, we show that it can greatly reduce human effort spent since continuous mapping redefinition on evolving ontologies is no longer necessary.