ONLI: An ontology-based system for querying DBpedia using natural language paradigm

• Current mechanisms for information retrieval from knowledge bases demands a high level of expertise.
• Natural language processing hides the formality of a knowledge base as well as the executable query language.
• We propose an ontology-based system for retrieving information from DBpedia.
• The system has been evaluated in order to known its accuracy on how to find the correct answer into DBpedia.

The Semantic Web has emerged as an extension of the current Web, in which Web content has well-defined meaning through the addition of logic-based metadata. However, current mechanisms for information retrieval from semantic knowledge bases restrict their use to only experienced users. To address this gap, the natural language processing (NLP) is deemed to be very intuitive from a use point of view, due to it hides the formality of a knowledge base as well as the executable query language. This paper presents a novel ontology-based information retrieval system for DBpedia called ONLI (Ontology-based Natural Language Interface). ONLI proposes the use of an ontology model in order to represent both the syntactic question’s structure and the question’s context. This model allows inferring the answer type expected by the user through an established question’s classification. These features allow reducing the search space thus increasing the probability of providing the correct answer. From this perspective, ONLI was evaluated in terms of their ability to find the correct answer into DBpedia’s content, achieving promising results and proving to be very useful to non-experienced users.