1. UKM Learning and Research Repository
2. PHD Thesis / Tesis PHD
3. Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat

Title:	Analysis of concept structure and semantic relations based on graph-independent structural analysis
Authors:	Maziar Amirhosseini (P49988)
Supervisor:	Juhana Salim, Prof. Dr.
Keywords:	Ontology evaluation Graph theory Structural domains Dissertations, Academic -- Malaysia Universiti Kebangsaan Malaysia -- Dissertations
Issue Date:	29-Jan-2016
Description:	Core literatures in ontology evaluation reveal that numerous methods and measures have been proposed to tackle problems which can be categorized into structural, functional and usability issues. Majority of the research in ontology evaluation have focused on functionality and usability rather than structural dimensions. To identify that two nodes in a graph are related, structural analysis in ontology has relied on graph theory. However, there are other areas in evaluating ontology structure such as the structure of concepts and the structural domains of semantic relations. These areas are called Graph-independent domains by researchers in this area of investigation which is a novel branch in structural analysis. This investigation proposed a novel method to evaluate concept structure and semantic relations’ domains as well as their interactive behaviour based on the graph-independent approach.This method is called OntoAbsolute. This method adapted Immanuel Kant’s knowledge theoryto achieve cognitive results to generate theoretical arguments in analyzing ontology structure. The adapted knowledge theory insists on the casual relations between simplicity in concepts structure and unity in semantic relations. There are two influential ontology evaluation methods, OntoClean and OntoMetric. OntoAbsolute have similarities and contrasts with OntoClean and OntoMetric, but this novel method quantitatively attempts to evaluate graph-independent structural domains trusted to knowledge theory. OntoAbsolute has revealed several research questions and hypotheses in analyzing structural domains. OntoAbsolute relied on a mixed-method which adapted the nine steps proposed by Houston to evaluatethe agricultural ontology entitled VocBench as the data resource. In this investigation, the sample size is 1500 concepts and the appropriate sampling technique is stratified random sampling. Data was recorded and analyzed by employing SPSS windows (release 19.0) and Excel software. Data analysis was divided into two complement steps; Firstly, research questions and general hypotheses were examined. Secondly, main hypotheseswere tested to achieve precise knowledge basis of mixed-method logical reasoning. The arguments’ results revealed a decrease in the numbers of complex concepts results in increasing simplicity in concepts structure. The size and domain of taxonomic and non-taxonomic relations have been realized in VocBench. The hypotheses’ outcomes generally expressed that increase of semantic relations depends on using simplest concepts as a casual factor. Lastly, the generated theory is that simplicity in concepts structure causes unity in semantic network. This generated theory results in achieving the highest level of cognition in graph-independent structural analysis in ontology verification and validation in improving IR performance.,Certification of Master's / Doctoral Thesis" is not available
Pages:	360
Call Number:	QA76.5913.A476 2016 3 tesis
Publisher:	UKM, Bangi
Appears in Collections:	Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.