SPARQL query answering with RDFS reasoning on correlated probabilistic data

  • Authors:

  • Chi-Cheong Szeto;Edward Hung;Yu Deng

  • Affiliations:

  • Department of Computing, The Hong Kong Polytechnic University, Hong Kong;Department of Computing, The Hong Kong Polytechnic University, Hong Kong;IBM T.J. Watson Research Center, Yorktown Heights, NY

  • Venue:
  • WAIM'11 Proceedings of the 12th international conference on Web-age information management
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

In recent years, probabilistic models for Resource Description Framework (RDF) and its extension RDF Schema (RDFS) have been proposed to encode probabilistic knowledge. The probabilistic knowledge encoded by these models ranges from statistical relationships over possible objects of an RDF triple to relationships among correlated triples. The types of queries executed on these models vary from single triple patterns to complex graph patterns written in SPARQL, a W3C query language for RDF. Some query answerings only include reasoning of transitive properties and others do not have any reasoning. In this paper, we propose answering SPARQL queries with RDFS reasoning on probabilistic models that encode statistical relationships among correlated triples. One result to note is that although uncertainties of explicitly declared triples are specified using point probabilities, the evaluation of answers involving derived triples results in interval probabilities. Moreover, we experimentally examine how the execution time of the proposed query answering scales with the data size and the percentage of probabilistic triples in the data.