Toward Automatic RDF Property Tagging

This work investigates some problems about semantic properties (also known as predicates) of Knowledge Bases, as part of the Semantic Web, for querying and ranking them toward a new system to tag automatically RDF Property over parts of free-text in Natural Language. The main insights and contributions are: - a contribution to develop a system called Qpedia, inspired by SWiPE, to make difficult query on schema-agnostic Knowledge Bases with a simple and intuitive mobile-user interface; - the creation of the first approach exploiting Machine-Learning to rank RDF predicates; - the creation of a possible approach to tagging free-text with RDF predicates, with a case study of possible backend; The proposed methods have been evaluated with the most popular Knowledge Bases (DBpedia, WikiData, MusicBrainz and Freebase), obtaining encouraging results. Thus, this work is a first step towards the RDF Property Tagging of natural language, as reflected in Chapter 5, needed to pave the way providing a resolution of sub-problems related to Question Answering over RDF properties, which are not typically addressed in literature through this way.