Question answering system for english translation of al-qur`an based on the logico-linguistic representational model

Abstract

Machines lack analytical and deduction capabilities to understand the gigantic reservoirs of data. The knowledge behind the data remains largely un-extracted. On the other hand, information retrieval mechanisms do not perform well in natural language environment. Search engines perform document-level analysis rather than content-level analysis, thus failing to understand the intent of the search and retrieving undesired results. The current indexing mechanism and retrieval strategies also warrant improvement. User intent is not taken into consideration while retrieving and presenting the final results from a search. Question answering systems help to find accurate and precise answers, with natural language as the mode of interaction. Building upon this recent trend, Muslim researchers have made numerous efforts for information retrieval systems for Al-Qur’an and Hadith since they constitute the foundation of Islam. Efforts have been made for knowledge representation and retrieval of Al-Qur’an to ease the process of procuring information. However, the approaches taken so far in this respect suffer from the same problems as other information retrieval systems. The current research aims at the development of knowledge representation formalism, rules and knowledge representation model for Al-Qur’an using combination of logic (First Order Predicate Logic) and linguistics (Semantics). It also aims to design a question answering system for Al-Qur’an that retrieves answers efficiently, accurately and precisely. This system is expected to understand the user intent before processing the queries and presenting the results, and uses natural language as the mode of interaction. It employs ADDIE (Analysis, Design, Development, Implementation and Evaluation) as the research methodology. ADDIE is well known for systematic problem solving and all the stages within it can be practiced interchangeably in the hierarchy. The results of the research clearly indicate that the mechanism of knowledge representation developed in this work is effective and provides intelligence to machines for deductive analysis of data. The system exhibits substantial improvement in results as compared to the previous approaches taken for knowledge representation and information retrieval. The major significance of this study is the ease it accords to the process of obtaining efficient, accurate and precise information from Al-Qur’an. Since the system interacts in natural language, it will be helpful for both experts of the domain knowledge as well as laymen. It will facilitate access of information for Muslims as well as Non-Muslims, thus contributing in spreading the message of Islam.