Butt, Anila Sahar2017-09-052017-09-05b45019757http://hdl.handle.net/1885/125224Ontology search, which is the process of finding ontologies or ontological terms for users’ defined queries from an ontology collection, is an important task to facilitate ontology reuse of ontology engineering. Ontology reuse is desired to avoid the tedious process of building an ontology from scratch and to limit the design of several competing ontologies that represent similar knowledge. Since many organisations in both the private and public sectors are publishing their data in RDF, they increasingly require to find or design ontologies for data annotation and/or integration. In general, there exist multiple ontologies representing a domain, therefore, finding the best matching ontologies or their terms is required to facilitate manual or dynamic ontology selection for both ontology design and data annotation. The ranking is a crucial component in the ontology retrieval process which aims at listing the ‘relevant0 ontologies or their terms as high as possible in the search results to reduce the human intervention. Most existing ontology ranking techniques inherit one or more information retrieval ranking parameter(s). They linearly combine the values of these parameters for each ontology to compute the relevance score against a user query and rank the results in descending order of the relevance score. A significant aspect of achieving an effective ontology ranking model is to develop novel metrics and dynamic techniques that can optimise the relevance score of the most relevant ontology for a user query. In this thesis, we present extensive research in ontology retrieval and ranking, where several research gaps in the existing literature are identified and addressed. First, we begin the thesis with a review of the literature and propose a taxonomy of Semantic Web data (i.e., ontologies and linked data) retrieval approaches. That allows us to identify potential research directions in the field. In the remainder of the thesis, we address several of the identified shortcomings in the ontology retrieval domain. We develop a framework for the empirical and comparative evaluation of different ontology ranking solutions, which has not been studied in the literature so far. Second, we propose an effective relationship-based concept retrieval framework and a concept ranking model through the use of learning to rank approach which addresses the limitation of the existing linear ranking models. Third, we propose RecOn, a framework that helps users in finding the best matching ontologies to a multi-keyword query. There the relevance score of an ontology to the query is computed by formulating and solving the ontology recommendation problem as a linear and an optimisation problem. Finally, the thesis also reports on an extensive comparative evaluation of our proposed solutions with several other state-of-the-art techniques using real-world ontologies. This thesis will be useful for researchers and practitioners interested in ontology search, for methods and performance benchmark on ranking approaches to ontology search.enOntology SearchOntology RankingRanking Model201610.25911/5d74e26f19b57