UMP Institutional Repository

Ontology based keyword suggestion for image processing domain

Ooi, Jessie (2018) Ontology based keyword suggestion for image processing domain. Masters thesis, Universiti Malaysia Pahang.

[img]
Preview
Pdf
Ontology based keyword suggestion for image processing.pdf - Accepted Version

Download (287kB) | Preview

Abstract

The ontology-based keyword search system is a system that will identify related keywords for a search session. The internet is becoming increasingly important day by day so does the search engines that help users to retrieve information from the internet. However, the query inefficiency is still a problem. This is often caused by the inappropriate keywords given by the users. A junior researcher often does not have the ability to come out with suitable keywords that can describe their information needs accurately due to the lack of basic knowledge in the domain. To improve the junior researcher’s searching experience, the query modification methods have been investigated. Based on this, an ontology-based keywords search system has been proposed. The development of the ontology-based keywords search system was divided into two different part. The first part of the development is the ontology development. Image processing has been selected as the domain for the ontology. The methodology used for the development of the ontology was based on the ontology development 101. The ontology developed were based on research articles and Image Processing reference books and the tools used during the ontology development is Protege. On the other hand, the proposed system was developed using SPARQL to query the information from image processing ontology. The evaluations of the ontology and the proposed system had been carried out using 4 different methods. The first evaluation used is the reasoning. This is to ensure there is no logical contradiction in the ontology. The second evaluation is the metric-based evaluation. The evaluation has shown that the ontology has a detailed type of knowledge representation and based on the AP of the metric-based evaluation, the ontology has the ability to present the knowledge available in the schema effectively. Furthermore, the Precision and Recall were calculated. The Precision and Recall rate of the keywords provided by the OKSS achieve a 0.78 precision rate which is 0.09 higher than the precision rate of the query suggestion provided by the Google search engine. Similarly, the query provided by the Google search engine has a lower average recall rate as compared to the OKSS’s keywords recall rate. To evaluate the usefulness (in terms of productivity, effectivity and time consumption) of the proposed approach, a user experiment has been conducted. In this study, only master students from the Faculty of Computer Systems and Software Engineering were selected. The system usability score had been calculated in this experiment where the proposed system has scored 81.62. The usability score has shown that the proposed system has achieved the study objectives.

Item Type: Thesis (Masters)
Additional Information: Thesis (Master of Science) -- Universiti Malaysia Pahang – 2018, SV: DR. MANSOOR ABDULLATEEF ABDULGABBER, NO. CD: 11547
Uncontrolled Keywords: Ontology; keyword; image processing
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Faculty/Division: Faculty of Computer System And Software Engineering
Institute of Postgraduate Studies
Depositing User: Mrs. Sufarini Mohd Sudin
Date Deposited: 14 Oct 2019 03:07
Last Modified: 18 Oct 2019 03:42
URI: http://umpir.ump.edu.my/id/eprint/26093
Download Statistic: View Download Statistics

Actions (login required)

View Item View Item