Using semantic approaches to answer Arabic questions from the Holy Quran

Other Title(s)

استخدام المنهجيات الدلالية لخدمة أنظمة الإجابة عن الأسئلة باللغة العربية من القرآن الكريم

Dissertant

Shmaisani, Hashim Adnan Jadan

Thesis advisor

Tartir, Samir

Comitee Members

al-Ghul, Said
al-Emam, Nimr
Sawalihah, Majdi

University

Philadelphia University

Faculty

Faculty of Information Technology

Department

Department of Computer Science

University Country

Jordan

Degree

Master

Degree Date

2014

English Abstract

-Arabic is the fifth language in the world according to number of speakers, and is considered a second language to over one billion Muslims worldwide.

Yet, very little research has been done on machine processing of Arabic text.

One of the main methods that is effected by language in computer science research areas, is the searching capability quality, Arabic language search is still in the phase of syntax based search, retrieving the exact match to the user based on the provided search keywords syntax, only some tried to move beyond that, but still no work was done regarding searched data structure or semantic Question Answering methodologies to serve this language.

Relying on the fact that in general languages are semantic based more than syntax based, a single word may give more than one meaning based on the context, it holds more meaning than just a string of characters.

Accordingly, we are proposing an approach to overcome the semantic and structural challenges that are currently limiting the quality of search among Arabic based content.

Mainly the goal is to solve semantic issues by using synonyms for the user query keywords, as for the content structure issues an ontology based approach is used in order to organize textual entities and their relationships with other content in the text, rather than simple strings of characters.

To test and implement this approach for the first time a very unique semantic based context is used, which is the Holy Quran.

Nowadays people searching the Holy Quran for knowledge are using term and exact-match based approaches, which is not suitable for a rich text as the Quran.

People searching Holy Quran are often looking for a concept among the rather than a certain text.

In other words, they are looking for the semantic similarity rather than the syntactic similarity.

Which explains the use of a semantic layer in our system.

Furthermore, the type of queries coming from the Holy Quran searchers are question based queries [e.g.

How many years did the people of the cave slept?], so in order to cover this XIII side a question analysis layer is embedded in order to detect keywords and understand users’ questions in order to determine the desired answer.

The used ontology content is based on the relation extractor model, which is a very important contribution to the Arabic ontology building.

The relation extracting process is the action of detecting and retrieving relations and their concepts among raw text.

This thesis is proposing a model that can detect relations and concepts among any given Arabic context depending on grammatical perspectives.

Which is, up to our knowledge, the first semantic relation extractor for the Arabic language.

In summary, this research builds a question answering layer over a semantic layer built over an ontology layer.

This will be implemented for the first time for the Arabic language

Main Subjects

Information Technology and Computer Science
Arabic language and Literature

Topics

• Data processing
• Semantics
• The holy koran
• Arabic language

No. of Pages

62

Table of Contents

• APA
• MLA
• AMA

Language

English

Data Type

Arab Theses

Record ID

BIM-546121