This paper extends the model theory of RDF with rules, placing an emphasis on integration with OWL and decidability of entailment. We start from an abstract syntax that views a rule as a pair of rule graphs which generalize RDF graphs by also allowing rule variables in subject, predicate and object positions. We include RDFS as well as a decidable part of OWL that weakens D-entailment and OWL Full. Classes can be used as instances. Almost all examples in the DAML set of test rules are covered by our approach. For a set of rules R, we define a general notion of R-entailment. Extending earlier results on RDFS and OWL, we prove a general completeness result for R-entailment. This result shows that a restricted form of application of rules that introduce blank nodes is sufficient to determine R-entailment. For rules that do not introduce blank nodes, we prove that R-entailment and R-consistency are decidable and in PSPACE. For rules that do not introduce blank nodes and that satisfy a bound on the size of rule bodies, we prove that R-consistency is in P, that R-entailment is in NP, and that R-entailment is in P if the target RDF graph is ground. © Springer-Verlag Berlin Heidelberg 2005.
CITATION STYLE
Ter Horst, H. J. (2005). Combining RDF and part of OWL with rules: Semantics, decidability, complexity. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 3729 LNCS, pp. 668–684). Springer Verlag. https://doi.org/10.1007/11574620_48
Mendeley helps you to discover research relevant for your work.