ResearchBib Share Your Research, Maximize Your Social Impacts
Sign for Notice Everyday Sign up >> Login

Towards an Automated Islamic Fatwa System: Survey, Dataset and Benchmarks

Journal: International Journal of Computer Science and Mobile Computing - IJCSMC (Vol.10, No. 4)

Publication Date:

Authors : ; ; ; ; ; ;

Page : 118-131

Keywords : Islamic Fatwa; Question Answering; Text Classification; Artificial Intelligence; Machine Learning; Deep Learning; Natural Language Processing;

Source : Downloadexternal Find it from : Google Scholarexternal

Abstract

Islam is the second largest and the fastest growing religion. The Islamic Law, Sharia, represents a profound component of the day-today lives of Muslims. This creates a lot of queries, about specific problems, that requires answers, or Fatwas. While sources of Sharia are available for anyone, it often requires a highly qualified person, the Mufti, to provide Fatwa. To get certified for Fatwa, the Mufti needs to undergo a sophisticated and long education process that starts from basic to high school. With Islam followers representing almost 25% of planet earth population, generating a lot of queries, and the sophistication of the Mufti qualification process, creating shortage in them, we have a supply-demand problem, calling for Automation solutions. This motivates the application of Artificial Intelligence (AI) to Automated Islamic Fatwa. In this work, we explore the potential of AI, Machine Learning and Deep Learning, with technologies like Natural Language Processing (NLP), paving the way to help the Automation of Islam Fatwa. We start by surveying the State-of-The Art (SoTA) of NLP, and explore the potential use-cases to solve the problems of Question answering and Text Classification in the Islamic Fatwa Automation. We present the first and major enabler component for AI application for Islamic Fatwa, the data. We build the largest dataset for Islamic Fatwa, spanning the widely used websites for Fatwa. Moreover, we present baseline systems, for Topic Classification, Topic Modelling and Retrieval-based Question-Answering, to set the direction for future research and benchmarking on our dataset. Finally, we release our dataset and baselines to the public domain, to help advance the future research in the area.

Last modified: 2021-04-27 21:51:45