Semantic keyword search for expert witness discovery

Abstract

In the last few years, there has been an increase in the amount of information stored in semantically enriched knowledge bases, represented in RDF format. These improve the accuracy of search results when the queries are semantically formal. However framing such queries is inappropriate for inexperience users because they require specialist knowledge of ontology and syntax. In this paper, we explore an approach that automates the process of converting a conventional keyword search into a semantically formal query in order to find an expert on a semantically enriched knowledge base. A case study on expert witness discovery for the resolution of a legal dispute is chosen as the domain of interest and a system named SKengine is implemented to illustrate the approach. As well as providing an easy user interface, our experiment shows that SKengine can retrieve expert witness information with higher precision and higher recall, compared with the other system, with the same interface, implemented by a vector model approach.