Ontology based natural language queries transformation into SPARQL queries

Abstract

Ontology-based Data Access (OBDA) enables semantic access to a set of heterogamous data sources, supporting data sharing, exchanging, and integration across these data sources. In the OBDA scheme, normally formal query languages, such as SPARQL, are used to represent the user questions, which limits end users from defining their requests. To cope with this problem a layer that accepts the user request in her own language and transforms it into one of these formal languages has become a necessity. To this end, we introduce a new and interactive method that guides the user during the translation. The proposed approach makes use of the capabilities of natural language processing and the semantic information embedded in the domain ontology. Furthermore, the proposed approach considers user involvement during the translation process. To demonstrate the effectiveness, we implemented the proposed approach and validated it against a query benchmark assessing the query accuracy and efficiency.