A Comparison of Unsupervised Taxonomical Relationship Induction Approaches for Building Ontology in RDF Resources

Conference paper

DOI: 10.1007/978-3-319-06826-8_32

Part of the Lecture Notes in Computer Science book series (LNCS, volume 8388)
Cite this paper as:
Zong N., Lee S., Kim HG. (2014) A Comparison of Unsupervised Taxonomical Relationship Induction Approaches for Building Ontology in RDF Resources. In: Kim W., Ding Y., Kim HG. (eds) Semantic Technology. JIST 2013. Lecture Notes in Computer Science, vol 8388. Springer, Cham

Abstract

Automatically generated ontology can describe the relationship of meta-data in Linked Data or other RDF resources generated from programs, and advances the utility of the data sets. Hierarchical document clustering methods used to generate concept hierarchies from retrieved documents or social tags can be used for constructing taxonomy or ontology for Linked Data and RDF documents. This paper introduces a framework for building an ontology using the hierarchical document clustering methods and compares the performance of three classic algorithms that are UPGMA, Subsumption, and EXT for building the ontology. The experiment shows EXT is the best algorithm to build the ontology for RDF resources and demonstrates that the quality of the ontology generated can be affected by the number of concepts that are used to represent the entities and to formalize the classes in the ontology.

Keywords

Taxonomical relationship induction Hierarchy generation Ontology generation RDF Linked data 

Copyright information

© Springer International Publishing Switzerland 2014

Authors and Affiliations

  1. 1.Biomedical Knowledge Engineering Lab, School of DentistrySeoul National UniversitySeoulKorea

Personalised recommendations