Evaluating Thesaurus Alignments for Semantic Interoperability in the Library Domain

Abstract

Published by the IEEE Computer Societyother controlled vocabularies) specific to fields, institutions, and even collections. The desire to make CH resources available to the general public (for example, see www. europeana.eu) increases the need to facilitate interoperability across different contexts.By providing representational standards such as the Simple Knowledge Organization System (SKOS; www.w3.org/2004/02/skos) and generic tool support, the Semantic Web community has taken a prominent role in this facilitation. Its ontology-matching branch aims at developing technology to produce alignments-that is, sets of semantic mappings bet ween elements from different vocabularies. 1 One can exploit alignments, for instance, to access a collection via thesauri it is not originally indexed with, to interconnect distributed, differently annotated collections on the object level, or to merge two thesauri to rationalize thesaurus maintenance.Unfortunately, our experience shows that existing matching tools often do not perform well in CH applications. 2 We believe part of the problem is that they strive for generality. To this end, we argue that the generation and evaluation of thesaurus alignments must take into account well-understood realworld application contexts and their specific requirements. Ontology MatchingOntology matching aims at determining the semantic relations bet ween elements of two given knowledge organization systems (for example, ontologies and thesauri). 1 The set of semantic relations usually comprises concept equivalence, hierarchical concept links

SCIF-IRIS framework: a framework to facilitate interoperability in supply chains

“…It involves the use of a technological platform that includes a thesaurus [18] and a set of ontologies, implemented using Protégé [19]. The difference between a thesaurus and an ontology is that the thesaurus does not contain any axioms that allow any kind of automatic reasoning to be performed.…”

A Web-Based Repository Service for Vocabularies and Alignments in the Cultural Heritage Domain

“…While the design of our repository services takes into account general needs of vocabulary experts, its current implementation responds to direct requirements of the STITCH project and its umbrella programme CATCH [4].…”

“…theeuropeanlibrary.org/telplus/) investigated methods to support metadata interoperability by automatically identifying inter-vocabulary semantic mappings [1,2]. They showed that the automatic matching of vocabularies can be the basis of various real-world usage scenarios, including support for indexing and re-indexing collection items, inter-collection search and navigation, but also thesaurus management [3,4]. First tools for vocabulary services have been deployed at the National Library of the Netherlands, so that there is indeed an industrial uptake of Semantic Web technology in this context.…”

“…A more expressive representation is required, however, to meet realistic application requirements such as the ones we encountered in CATCH [4]. Here, mappings need to have properties of their own, for instance, "confidence measure", "producer", or "evaluated as".…”

Vocabulary Matching for Book Indexing Suggestion in Linked Libraries – A Prototype Implementation and Evaluation

“…Many of these books (nearly 238K) are thus indexed with both Biblion and Brinkman. Following the reindexing evaluation approach of [11], we used the existing Brinkman subjects of these books as a gold standard with which we could automatically compare new suggested subjects. Based on the low performance of this preliminary automatic evaluation, we decided to drop these first suggestion rules.…”