Dissertation/Thesis Abstract

A Framework for Building Lightweight Ontologies Based on Semi-Structured Data for Semantic Annotation
by Ali, Elshaimaa Elsayed, Ph.D., University of Louisiana at Lafayette, 2015, 147; 10188059
Abstract (Summary)

The Semantic Web vision was created in 1999 by Tim Berners Lee. In 2001, The W3C declared the Semantic Web as the web of shared data. Hence, the realization of this vision requires linking definitions of concepts in web documents to a semantically integrated and interoperable structure to define a domain. This process is known as semantic annotation. However, the main challenge in the annotation process is the creation of domain ontologies.

This dissertation is proposing a framework for building lightweight domain ontologies, as instances of an interoperable semantic structure. We designed a multilayered interoperable semantic schema, to define lightweight domain ontologies. This is an extension of SKOS schema, and makes use of the Wikipedia structure for defining concepts. Also, we proposed a Wikipedia-based approach for extracting concepts for ontology learning.

We developed the experimental work using a prepared set of clustered concepts in a domain of interest, which is the Information Retrieval domain. We attained almost 85% accuracy in terms of the Rand Index and a precision value of about 45%. Then we used SPARQL queries, to extract relevant concept properties from DBpedia, and map it to the definition of the modeled annotation schema. The produced triples demonstrated a promising direction to explore for generating lightweight ontologies to be used for annotation. This is considered an important step towards the realization of the Linked Data vision of the Semantic Web.

Indexing (document details)
Advisor: Raghavan, Vijay V.
Commitee: Benton, Ryan, Loganantharaj, Rasiah, Maida, Anthony
School: University of Louisiana at Lafayette
Department: Computer Science
School Location: United States -- Louisiana
Source: DAI-B 79/01(E), Dissertation Abstracts International
Source Type: DISSERTATION
Subjects: Computer science
Keywords: Concept extraction, Ontology, Semantic modeling, Semantic web, Wikipedia
Publication Number: 10188059
ISBN: 9780355112665
Copyright © 2019 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy
ProQuest