Visualisation of Relatedness

Abstract

The knowledge modelling potential of the Semantic Web technologies and the ways that the latter can be utilised assist and stimulate knowledge exchange is investigated. The best option from the available online knowledge repositories developed by the online community is sought. It transpires that DBpedia is the most strongly linked data set within the Web of Data and that its knowledge base covers a wide range of different domains, representing the time-evolving conceptual collaborative agreements of authors of Wikipedia articles. Existing DBpedia applications are identified and analysed e.g. DBpedia Spotlight and DBpedia RelFinder. None of the identified DBpedia applications instantiate all the steps constituting the process-based semantic exploration search. In this work, a system that implements the entire process mentioned is developed. It provides interactive visualisation for semantically-based exploration of Wikipedia content. When the user of the system submits a text query, a semantic graph consisting of relevantly labelled nodes will appear on the screen. The layout of the graph can be rearranged by the user. The tool box provides for zooming, panning and overview of the visualisation. The nodes can be clicked for further semantic expansion. The level of semantic expansion of the nodes is limited only by the existence of relevant DBpedia knowledge. The system, dubbed Visualisation of Relatedness, was evaluated through an online survey that included a questionnaire to establish the opinion of the users as regards ease of use, navigation, interactivity, clarity of visualisation, usability and usefulness of the system. Most users rated the system highly in all aspects. In particular, they found the system to be useful in discovering new information within Wikipedia content. This implies that the system facilitates knowledge acquisition and the subsequent exchange of the same.