Exploiting web resources for the identification of relations between concepts: a Java based implementation and case study

A Java based implementation of an Ontology Evolution Manager was described in Gabbanini (2010): it is a framework offering a set of tools to support processes of manipulation and growth of ontological knowledge bases, based on inputs consisting in free text documents. The Ontology Evolution Manager can be used to support the process of Ontology Evolution, i.e., the process of identifying potential novel entities and relationships to be included in an established ontology. This report describes a Java based application, built using the Ontology Evolution Manager, intended to perform ontology evolution processes, by enriching ontologies with new relations. The enrichment phase uses as sources of background knowledge the WordNet repository (see WordNet, 2010) and the Scarlet system (Sabou et al., 2008, Sabou et al., 2008b, Scarlet, 2010). The application is based on ideas described in Zablith et al. (2009), but new ideas have been introduced and the code has been implemented from scratch by the author, so as to be reusable within the framework of the Collective Knowledge Management System described by Burzagli et al. (2010). The report describes techniques and implementation details, along with a test case in which an ontology, built within the e-Inclusion Laboratory1 to describe the domain of inclusive tourism, is enriched with entities and relationships generated from the analysis of textual reviews, contributed by customers of a real web based service that allows booking and commenting on the accessibility of a selection of accommodation resources all over the world.