REWOrD: Semantic Relatedness in the Web of Data

This paper presents REWOrD, an approach to compute semantic relatedness between entities in the Web of Data representing real word concepts. REWOrD exploits graph nature RDF data and SPARQL query language access this data. Through simple queries, constructs weighted vectors keeping informativeness predicates used make statements about being compared. The most informative path is also considered further refine informativeness. Relatedness then computed by cosine vectors. Differently from previous approaches based on Wikipedia, does not require any prepro- cessing or custom transformation. Indeed, it can lever- age whatever knowledge base as a source background knowledge. We evaluated different settings using new dataset investigate its flexibility. As compared related work classical datasets, obtains comparable results while, one side, avoids burden preprocessing transformation and, other provides more flexibility applicability broad range domains.