Benchmarking RDF Metadata Representations: Reification, Singleton Property and RDF*

摘要

RDF reification is a data modelling solution for writing RDF statements about RDF statements. A set of approaches exist in the literature to express statement-level metadata, or “reified” statements, in RDF. They have primarily been designed to attach additional contextual information to individual triples, such as provenance, spatio-temporal validity, or certainty. Practically, different methods have been developed to present such metadata while complying with the RDF standard and syntax. However, when effectively stored in triplestores, these various solutions produce diverse results in terms of querying performance, storage efficiency and usability. In this article, we analyse these metrics across three popular reification approaches: Standard Reification, Singleton Property and the recent RDF*. We publish our benchmark evaluation resources so that the community could expand and compare the study on different datasets, reification approaches and storage implementations.