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

Title:	Hair oriented data model for spatio-temporal data mining
Authors:	Shahd Abbas Madraky (P53568)
Supervisor:	Abdul Razak Hamdan, Prof. Dr.
Keywords:	Data mining. Hair-Oriented Data Model
Issue Date:	28-Jul-2015
Description:	Having an effective Decision Support System (DSS) by mining regards to fast data changing is one of the most important demands in spatio-temporal (ST) data. The satellite images, weather maps, transportation information are some examples of spatiotemporal data. The spatio-temporal data types are complex in terms of numerous attributes, continuously value changing, and the data increasing towards time. There are several methodologies, frameworks, models and algorithms have been proposed to represent and process the spatio-temporal data, either utilizing the entity relationship, object-oriented and object-relational data model. However, the nature of complexity of spatio and temporal data decreases the effectively of current models significantly. In order to address data complexity and having a better object-relational data model in spatio-temporal DSS systems the operational and mining properties are improved in this research. Therefore, this study aims to propose a Hair-Oriented Data Model (HODM) suitable for spatio-temporal data inspired by the nature of hair growth based on its structure has root and shaft. The concepts, data structure and operations for the HODM are represented throughout mathematical formulae. The HODM is exploited to store, retrieve, process and analyze the data. A data definition and manipulation language for HODM has been proposed include three kind of operations are: data definition/manipulating operations (such as Grow, Cut, Implant, Fall, Wash and Perm), data mining operations (such as Comb, Plait and Color) for clustering and classification, and data preserving operations (such as Cover, Tangle and Wig). The proposed data model is developed using object-relational approach in Oracle and the UCI climate change spatio-temporal datasets have been used to measure the HODM efficiency. The efficiency of HODM is evaluated by reducing the data redundancy, the query response time and executing time for data mining and improving mining accuracy. In this regard, this study has also shown the development process of a climate change spatio-temporal model using k-mean, support vector machine and regression model techniques using the proposed data mining operations been proposed. The significances of the work are elaborated by doing comparison and statistical analysis with the best spatio-temporal data model namely Modeling of Application Data as MADS and Geoscience. The experimental results showed that the proposed data model for spatio-temporal data mining is easier to develop and more efficient.,Ph.D
Pages:	291
Call Number:	QA76.9.D343 S533 2015 3
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.