KASHAF: A KNOWLEDGE-GRAPHS APPROACH SEARCH-ENGINE FOR HADITH ANALYSIS & FLOW-VISUALIZATION

Abstract

Hadith is the term used to describe the narration of the sayings and actions of Prophet Mohammad (p.b.u.h.). The process of Hadith Inquiry can be modeled into a pipeline of tasks performed on a collection of textual data. Although tasks from the problem space of the digital Hadith domain match known data mining applications such as retrieval and visualization, existing solutions are repetitive, limited, and mainly manually constructed. In this research, I define 13 tasks found in Hadith studying problem space, survey and compare seven available Hadith data sources, and introduce KASHAF, a visual Hadith search engine that presents interactive flow-visualizations for narrative analysis of more than 60,000 Hadiths from the Nine-Books collection. To handle complex relational expressions about Hadith knowledge, I reformulate the Hadith retrieval problem as a two-fold solution: robust knowledge-graph querying; and Takhreej groups retrieving using semantic-similarity classification. The classifier was built by fine-tuning an AraBERT transformer model on a 200k pairs sample with an 80-20 split ratio and scored 0.9 recall and precision. This work demonstrated how the Hadith Inquiry Process could be more efficient and insightful with a user-centered methodology. Further, this research explores the gaps in the literature and displays the great potential of future computational Hadith research.

Date of Award	2021
Original language	American English
Awarding Institution	HBKU College of Science and Engineering