A framework for analysing the complexity of ontology

Show simple item record

dc.contributor.author Kazadi, Yannick Kazela
dc.date.accessioned 2021-06-15T03:12:01Z
dc.date.available 2021-06-15T03:12:01Z
dc.date.issued 2016-11
dc.identifier.uri http://hdl.handle.net/10352/456
dc.description M. Tech. (Department of Information and Communication Technology, Faculty of Applied and Computer Sciences), Vaal University of Technology en_US
dc.description.abstract The emergence of the Semantic Web has resulted in more and more large-scale ontologies being developed in real-world applications to represent and integrate knowledge and data in various domains. This has given rise to the problem of selection of the appropriate ontology for reuse, among the set of ontologies describing a domain. To address such problem, it is argued that the evaluation of the complexity of ontologies of a domain can assist in determining the suitable ontologies for the purpose of reuse. This study investigates existing metrics for measuring the design complexity of ontologies and implements these metrics in a framework that provides a stepwise process for evaluating the complexity of ontologies of a knowledge domain. The implementation of the framework goes through a certain number of phases including the: (1) download of 100 Biomedical ontologies from the BioPortal repository to constitute the dataset, (2) the design of a set of algorithms to compute the complexity metrics of the ontologies in the dataset including the depth of inheritance (DIP), size of the vocabulary (SOV), entropy of ontology graphs (EOG), average part length (APL) and average number of paths per class (ANP), the tree impurity (TIP), relationship richness (RR) and class richness (CR), (3) ranking of the ontologies in the dataset through the aggregation of their complexity metrics using 5 Multi-attributes Decision Making (MADM) methods, namely, Weighted Sum Method (WSM), Weighted Product Method (WPM), Technique for Order Preference by Similarity to Ideal Solution (TOPSIS), Weighted Linear Combination Ranking Technique (WLCRT) and Elimination and Choice Translating Reality (ELECTRE) and (4) validation of the framework through the summary of the results of the previous phases and analysis of their impact on the issues of selection and reuse of the biomedical ontologies in the dataset. The ranking results of the study constitute important guidelines for the selection and reuse of biomedical ontologies in the dataset. Although the proposed framework in this study has been applied in the biomedical domain, it could be applied in any other domain of Semantic Web to analyze the complexity of ontologies. en_US
dc.publisher Vaal University of Technology en_US
dc.subject Complexity of ontology en_US
dc.subject Semantic Web en_US
dc.subject.lcsh Dissertations, Academic -- South Africa en_US
dc.subject.lcsh Semantic web en_US
dc.subject.lcsh Ontologies (Information retrieval) en_US
dc.title A framework for analysing the complexity of ontology en_US
dc.type Thesis en_US
dc.contributor.supervisor Fonou-Dombeu, Dr. J. V.
dc.contributor.co-supervisor Okosun, Prof. K.

Files in this item

This item appears in the following Collection(s)

Show simple item record

Search DigiResearch

Advanced Search

My Account