Accurate Identification of Ontology Alignments at Different Granularity Levels

Xiaocao Hu, Zhiyong Feng, Shizhan Chen, Keman Huang, Jianqiang Li, Mengchu Zhou

Research output: Contribution to journalArticlepeer-review

10 Scopus citations

Abstract

As more and more ontologies are defined with different terms, ontology matching plays a crucial role in addressing the semantic heterogeneity problem in many different disciplines. Many efforts have been made to discover correspondences among terms in different ontologies. Most studies directly match two ontologies by utilizing terminological and structural methods that rely on ontologies themselves only. However, the decentralized characteristic of ontologies raises the uncertainty in ontology matching. To address this problem, we propose a four-stage ontology matching framework (FOMF) to enhance ontology matching performance. It is built upon the commonly accepted claim that an external comprehensive knowledge base can be used as a semantic bridge between domain ontologies for ontology matching. First, FOMF semantically maps domain ontologies to a knowledge base and then produces different types of alignments, including equivalence, subclass, sameas, and instance alignments. Similarities between two domain ontologies are next employed to enhance the equivalence and sameas alignments discovery. Finally, based on acquired alignments, inferred alignments are deduced to guarantee the completeness of matching results. Our experimental results show the superiority of the proposed method over the existing ones.

Original languageEnglish (US)
Article number7581070
Pages (from-to)105-120
Number of pages16
JournalIEEE Access
Volume5
DOIs
StatePublished - 2017

All Science Journal Classification (ASJC) codes

  • Computer Science(all)
  • Materials Science(all)
  • Engineering(all)

Keywords

  • Knowledge base
  • Ontology alignment
  • Ontology matching
  • Semantic heterogeneity

Fingerprint

Dive into the research topics of 'Accurate Identification of Ontology Alignments at Different Granularity Levels'. Together they form a unique fingerprint.

Cite this