Natural language query interpretation into SPARQL using patterns

Abstract

Our purpose is to provide end-users with a means to query ontology based knowledge bases using natural language queries and thus hide the complexity of formulating a query expressed in a graph query language such as SPARQL. The main originality of our approach lies in the use of query patterns. In this article we justify the postulate supporting our work which claims that queries issued by real life end-users are variations of a few typical query families. We also explain how our approach is designed to be adaptable to different user languages. Evaluations on the QALD-3 data set have shown the relevancy of the approach.