Preserving Information Content in RDF using Bounded Homomorphisms

Presented at: 9th Extended Semantic Web Conference (ESWC2012)

by Audun Stolpe, Martin G. Skjæveland

The topic of study in the present paper is the class of RDF homomorphisms that substitute one predicate for another throughout an set of RDF triples, under the condition that the predicate in question is not also a subject or object. These maps turn out to be suitable for reasoning about similarities in information content between two or more RDF graphs. As such they are very useful e.g. for migrating data from one RDF vocabulary to another. In this paper we address a particular instance of this problem and try to provide an answer to the question of when we are licensed to say that data is being transformed, reused or merged in a non-distortive manner. We place this problem in the context of RDF and Linked Data, and study the problem in relation to SPARQL construct queries.

Keywords: Data Transformation, Homomorphisms, RDF, SPARQL


Resource URI on the dog food server: http://data.semanticweb.org/conference/eswc/2012/paper/research/136


Explore this resource elsewhere: