Evaluation of Expressing without asserting approaches in RDF. The case of Conjectures

In this paper, we evaluate the existing reification approaches for expressing without asserting (EWA) statements in RDF along with their related contextual information and compare them with a new method, called Conjectures. Conjectures express RDF statements with three states of knowledge: undisputed claims, disputed claims, and settled claims. Conjectures extend the semantics of RDF Named graphs and introduce a new syntactical form to represent both conjectural and asserted information. Our evaluation tests were performed on a large sample of Wikidata entities about artworks interspersed with additional dummy statements to simulate alternative or abandoned claims and enrich the set of non-asserted claims. Our study evaluates metrics such as the total number of triples, loading time, dataset weight, and in particular query execution time for many different and meaningful types of queries. Results show that Conjectures is competitive with existing methods and outperforms other methods in terms of efficiency when retrieving debated statements, thereby demonstrating its potential as an effective tool for expressing nuanced RDF statements.