Paper 103 (Research track)

Detecting Erroneous Identity Links on the Web using Network Metrics

Author(s): Joe Raad, Wouter Beek, Frank Van Harmelen, Nathalie Pernelle, Fatiha Saïs

Abstract: Although best practices for publishing Linked Data encourage the re-use of existing IRIs, we often find ourselves with several resources denoting the same thing, requiring the use of owl:sameAs to align different identifiers on the Web. However, several studies dating as far back as 2010 have observed multiple misuses of this equality link, resulting in many erroneous and even inconsistent statements. In this paper, we show how network metrics such as the communities structure of the owl:sameAs graph can be used to detect these possibly erroneous statements. We have evaluated our methods on a subset of the LOD cloud, containing over 550M owl:sameAs statements.

Keywords: linked open data; identity; owl:sameAs; communities

Share on

Leave a Reply

Your email address will not be published. Required fields are marked *