Reasoning in the Semantic Web (SW) commonly uses Description Logics (DL) via OWL2 DL ontologies, or SWRL for variables and Horn clauses. The Rule Interchange Format (RIF) offers more expressive rules but is defined outside RDF and rarely adopted. For querying, SPARQL is a well-established standard operating directly on RDF triples. We leverage SPARQL construct queries as logic rules, enabling (1) an expressive, familiar SW rule language, and (2) general recursion, where queries can act on the results of others. We translate these queries to the Notation3 Logic (N3) rule language, allowing use of existing reasoning machinery with forward and backward chaining. Targeting a one-to-one query–rule mapping improves exchangeability and interpretability. Benchmarks indicate competitive performance, aiming to advance the potential of rule-based reasoning in the SW.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

SPARQL in N3: SPARQL construct as a Rule Language for the Semantic Web

  • Dörthe Arndt,
  • William Van Woensel,
  • Dominik Tomaszuk

摘要

Reasoning in the Semantic Web (SW) commonly uses Description Logics (DL) via OWL2 DL ontologies, or SWRL for variables and Horn clauses. The Rule Interchange Format (RIF) offers more expressive rules but is defined outside RDF and rarely adopted. For querying, SPARQL is a well-established standard operating directly on RDF triples. We leverage SPARQL construct queries as logic rules, enabling (1) an expressive, familiar SW rule language, and (2) general recursion, where queries can act on the results of others. We translate these queries to the Notation3 Logic (N3) rule language, allowing use of existing reasoning machinery with forward and backward chaining. Targeting a one-to-one query–rule mapping improves exchangeability and interpretability. Benchmarks indicate competitive performance, aiming to advance the potential of rule-based reasoning in the SW.