Developing an Ontology for Improving Question Answering in the Agricultural Domain

Numerous resources have been developed to have a better access to scientific information in the agricultural domain. However, they are rather concerned with providing general metadata of bibliographic references, which prevents users from accessing precise agricultural information in a transparent and simple manner. To overcome this drawback, in this paper, we propose to use domain-specific resources to improve the results in the answers obtained by an Open-Domain Question Answering (QA) system, obtaining a QA system for the agricultural domain. Specifically, it has been made by (i) creating an ontology that covers concepts and relationships from journal publications of the agricultural domain, (ii) enriching this ontology with some public data sources (e.g the Agrovoc thesaurus and the WordNet lexical database) in order to be precisely used in an agricultural domain, and (iii) aligning this enriched ontology with articles from our case-study journal, i.e. the Cuban Journal of Agricultural Science. Finally, we have developed a set of experiments in order to show the usefulness of our approach.