Fusion - Visually exploring and eliciting relationships in linked data

Abstract

Building applications over Linked Data often requires a mapping between the application model and the ontology underlying the source dataset in the Linked Data cloud. This mapping can be defined in many ways. For instance, by describing the application model as a view over the source dataset, by giving mappings in the form of dependencies between the two datasets, or by inference rules that infer the application model from the source dataset. Explicitly formulating these mappings demands a comprehensive understanding of the underlying schemas (RDF ontologies) of the source and target datasets. This task can be supported by integrating the process of schema exploration into the mapping process and help the application designer with finding the implicit relationships that she wants to map. This paper describes Fusion - a framework for closing the gap between the application model and the underlying ontologies in the Linked Data cloud. Fusion simplifies the definition of mappings by providing a visual user interface that integrates the exploratory process and the mapping process. Its architecture allows the creation of new applications through the extension of existing Linked Data with additional data. © 2010 Springer-Verlag.