Linked Open Ontology cloud: managing a system of interlinked cross-domain lightweight ontologies

Abstract

Metadata about documents, artifacts, and other objects is traditionally created by filling in metadata fields (e.g., dc:subject) with values taken from controlled vocabularies, such as keyword thesauri and classifications (e.g., LCSH). When this practice is applied to Linked Data a new problem is encountered: Linked data typically comes from different organizations and domains where mutually incompatible thesauri and classifications are used in annotations. This breaks links between the annotations, which creates data silos in linked data clouds. This paper argues that to solve the problem it is not enough to link data using primarily owl:sameAs mappings, as is customary today in, e.g., the Linked Open Data cloud, but one has to link annotation vocabularies, i.e., ontologies, into a Linked Open Ontology cloud. Since class hierarchies (using, e.g., rdfs:subClassOf) form the backbone of annotation ontologies, a key problem here is how to create and maintain a system of interlinked hierarchical ontologies so that the transitive subclass relations are not broken, when reasoning across ontologies in fundamental tasks such as query expansion and property inheritance. As a solution, we present the steps necessary for transforming thesauri into a cloud of ontologies and maintaining the system when ontologies are updated. Our approach has been used and evaluated in practice in building a cloud called KOKO of sixteen ontologies, with a total of 47,000 concepts, forming a basis for the Finnish national Linked Data architecture. KOKO has been published as an ontology service and is in use in, e.g., collection managing systems for both data indexing and semantic search.

Digital Heritage. Progress in Cultural Heritage: Documentation, Preservation, and Protection

“…Maintenance of Linked Data involves ontology evolution and Linked Data Dynamics, which have been studied in previous research in different contexts [131,156,224,192,131,64,164,8,101,204,136,169]. This thesis improves the state of the art by providing the observed change propagation scenarios with a proposal for a practical solution in the case of a centrally managed Linked Data Cloud.…”

“…Changes in an ontology may need to be propagated in three different scenarios: 1) inside the ontology, 2) to instances in a dataset using the ontology, and 3) to depending ontologies and applications [192]. A variety of infrastructures [131,64] and tools [225,164] have been proposed for handling ontology evolution. The same challenge of change propagation is evident with all of linked open data [8,101,204,136,169], often called Linked Data Dynamics in this context.…”

Linked Vocabularies for Mobility and Transport Research

“…Another project that focused on making several domain-specific legacy thesauri fit for the semantic web was FinnONTO. Here, it was not only attempted to align a vocabulary with others but to integrate thesauri describing different domains into a single resource ( KOKO ) [ 22 , 23 ]. The creation of KOKO is comparable to that of the Linked Vocabularies for SIS move, e.g.…”

Maintaining a Linked Data Cloud and Data Service for Second World War History

“…Early works on this line of research include, e.g., [11,15]. In [20,3], the problem of managing a set of interlinked hierarchical RDFS thesauri is discussed. Ontology evolution, and the propagation of changes caused by it, has been discussed in [23] and [25].…”