Semantic links across distributed heterogeneous data

Abstract

Data is often distributed using different formats and terminology to describe the same entities. This can make it difficult to query across data, to find duplicates, and to be aware of related resources. One approach to these issues is unifying the data into a shared format, connecting related entities. We created a prototype human aided system that converts data to RDF and links related entities using semantic web technologies. By using federation, entity resolution, and ontology alignment, common ideas are linked and the user can query across distributed data sources. In this paper, we show that by using our system, the time required to meaningfully connect data is significantly decreased compared to manual linking, showing an improvement of 78% with validation and 99% without validation on an inventory dataset with 11 thousand triples.