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

Title:	Application of lexical ontology for semi-automatic of logical data decision in data warehouse
Authors:	Mior Nasir Mior Nazri (P35238)
Supervisor:	Shahrul Azman Mohd Noah, Prof. Dr.
Keywords:	Database management Natural language processing (Computer science) Data warehousing Universiti Kebangsaan Malaysia -- Dissertations Dissertations, Academic -- Malaysia
Issue Date:	9-Jul-2012
Description:	Designing a data warehouse based on the current operational database is a very complex and time consuming process. The most critical part in designing a data warehouse is to identify the suitable entities as fact tables. Many researches incorporate artificial intelligence algorithm in the form of knowledge-based systems in order to assist in designing process. Most of existing data warehouse design tools employ direct transformation of input into corresponding designs and rely on the users to identify suitable entities to be modelled as fact tables. As a result, the main task in knowledge intensive model design still relies heavily on the user input. Hence, the main objective of this research is to incorporate knowledge-based system for the task of designing multidimensional model for data warehouse. We propose a new method based on supply driven approach using lexical ontology as the knowledge domain. Our method is able to make intelligent decision in designing data warehouse model by extracting valuable information from the knowledge domain. Once fact table is identified, the following step is to generate data warehouse multidimensional model with minimal user intervention. The feasibility of the proposed method is demonstrated by a prototype called ADW-tool using WordNet as knowledge domain. The process starts with the conversion of an enterprise logical model into a specification language model as input. The input goes through a set of synthesis and diagnosis rules before it is accepted into the system internal tables. The next stage involves identifying fact table and dimensional table candidates with the help from WordNet database. In the final stage, a logical multidimensional model in the form of Star schema emerged around the selected fact tables. Two sets of tests are performed by using selected business enterprise logical model as input. The two schemas used as input are identical with the input schemas used by two previous researches. Hence the result from this experiment is comparable against the result from those mentioned research defined as the bench mark. The research has demonstrated the viability of exploiting ontology in assisting the process of semi-automated data warehouse design for selecting potential facts and dimensional tables. Ontology is not claimed to be a panacea, however exploiting lightweight ontologies such as WordNet is seen able to suggest the correct entities for potential fact tables which will remain unidentified for novice human designer.,Certification of Master's / Doctoral Thesis" is not available
Pages:	274
Call Number:	QA76.9.D37M537 2012 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.