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

Title:	Islamic concept relations extraction using lexico-syntactic pattern approach
Authors:	Ammar Abdulateef Ali Al-Rawi (P74138)
Supervisor:	Saidah Saad, Dr.
Keywords:	Query processing Concept Hierarchy Taxonomy Islamic domain Lexico-syntactic pattern Dissertations, Academic -- Malaysia
Issue Date:	5-Sep-2016
Description:	A machine readable dictionary (MRD) is an electronic dictionary that enables query processing. One of the common processing tasks that has been widely applied is Concept Hierarchy which aims at identifying concepts with its corresponding taxonomies such as named entities, synonyms and hyponyms. The Islamic domain contains a variety of concepts that are associated with numerous taxonomies. Few research efforts have addressed the concept of hierarchy from the Islamic domain. Such efforts have utilized basic patterns and rules such as 'part of', 'such as' and 'is-a'. However, the Islamic domain contains more complicated concepts and taxonomies which makes the process of concept hierarchy from such a domain a challenging task. Therefore, this study aims to propose an concept hierarchy for the Islamic domain by extending the patterns and rules. In fact, the proposed patterns and rules extension aims to utilize lexico-syntactic patterns. The Islamic dictionary-glossary dataset used in this study was collected from the DEED International Islamic University of Malaysia website. A pre-processing task was applied by splitting sentences in order to facilitate the process of extracting definitions. In addition, Term Frequency-Inverse Document Frequency (TF-IDF) was carried out in order to identify the most frequently used concepts. Furthermore, two syntactical features were used including POS tagging and chunk parser in order to identify the tagging for each word (e.g. verb, noun, adjective, etc.) and extracting Noun Phrases (NP). In this manner, multiple n-gram methods were used including unigram, bi-gram and tri-gram. The evaluation was performed using precision method by identifying the number of correctly extracted concepts and relation between them. Moreover, an expert review evaluation was performed by an expert in the Islamic domain. The experimental results showed that the proposed method achieved 82% precision. That demonstrates the usefulness of extending rules for the Islamic domain.,Certification of Master's/Doctoral Thesis" is not available
Pages:	72
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.