Ontology evaluation approach for semantic web documents

• Main Author: Mohd. Hamka, Nurhamizah
• Format: Thesis
• Language: English
• Published: 2015
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://eprints.utm.my/id/eprint/77990/1/NurhamizahMohdHamkaMFC20151.pdf; http://eprints.utm.my/id/eprint/77990/; http://dms.library.utm.my:8080/vital/access/manager/Repository/vital:90326
• Institution: Universiti Teknologi Malaysia

Ontology is a conceptual tool used for managing and capturing information related to domain knowledge, such as the travel, education and medical domains. Publicly available ontology repositories like Falcons and SWOOGLE enhance the growth of ontology on the Web by providing a medium for ontology developers to publish their ontologies. In order to promote ontology reuse, a suitable approach for ontology evaluation is required that deals with ontology coverage for domain representation which includes an approach for validating the ontology with a corpus of information containing terms related to domain knowledge. Since contributions in ontology evaluation were introduced in different aspects, it is important to conceptualise related information to build an evaluation approach that can help users to select ontology. This work proposed OntoUji, an ontology that conceptualises information related to ontology evaluation. From OntoUji conceptualisation, these works proceed with the development of evaluation steps that are then converted into ontology evaluation algorithms to evaluate ontology documents retrieved from selected repositories according to data-driven evaluation approach. The data-driven approach focuses on evaluating the coverage of ontology using a set of keywords provided, yet similarly involves a comparison of ontological vocabulary with a pre-defined corpus, WordNet, gained from the information retrieval approach. The evaluation is then processed using Letters Pair Similarity algorithm as the selected similarity measures technique to process the ontology coverage result. The findings showed that the OntoUji ontology conceptualization helps to define ontology evaluation steps to gain similarity result for ontology selection.