Evaluating a Semantic Approach to Address Data Interoperability

Abstract

Semantic approaches have been used to facilitate interoperability in different fields of study. Current literature, however, shows that the semantic approach has not been used to facilitate the interoperability of addresses across domains. Addresses are important reference data used to identify locations and /or delivery points. Interoperability of address data across address or application domains is important because it facilitates the sharing of address data, addressing software and tools which can be used across domains. The aim of this research study has been to evaluate how a semantic (ontologies) approach could be used to facilitate address data interoperability and what the challenges and benefits of the semantic approach are. To realize the hypothesis and answer the research problems, a multi-tier hierarchy of ontology architecture was designed to integrate (across domain) address data with different levels of granularities. Four-tier hierarchy of ontologies was argued to be the optimal architecture for address data interoperability. At the top of the hierarchy was Foundation-Tier that includes vocabularies for location-related information and semantic language rules and concepts. The second tier has address reference ontology (called Base Address Ontology) that was developed to facilitate interoperability across the address domains. Developing optimal address reference ontology was one of the major goals of the research. Different domain ontologies were developed at the third tier of the hierarchy. Domain ontologies extend the vocabulary of the BAO (address reference ontology) with domain specific concepts. At the bottom of the hierarchy are application ontologies that are designed for specific purpose within an address domain or domains. Multiple scenarios of address data usage were considered to answer the research questions from different perspectives. Two interoperable address systems were developed to demonstrate the proof of concepts for the semantic approach. These interoperable environments were created using the UKdata+UPUdata ontology and UKpostal ontology, which illustrate different use cases of ontologies that facilitate interoperability. Ontology reason, inference, and SPARQL query tools were used to share, exchange, and process address data across address domains. Ontology inferences were done to exchange address data attributes between the UK administrative address data and UK postal service address data systems in the UKdata+UPUdata ontology. SPARQL queries were, furthermore, run to extract and process information from different perspective of an address domain and from combined perspectives of two (UK administrative and UK postal) address domains. The second interoperable system (UKpostal ontology) illustrated the use of ontology inference tools to share address data between two address data systems that provide different perspectives of a domain.