QURANIC KNOWLEDGE GRAPH: A MULTIDISCIPLINARY APPROACH TO MAPPING SEMANTIC NETWORKS IN THE QURAN

Understanding the Al-Quran is the eternal goal of every Muslim, wherever they are. Before the emergence of the Knowledge Graph, readers of the Al-Quran needed to move from the Al-Quran to Tafsir, Hadith, scholar opinions or other resources in order to grasp the full meaning of a verse or a chapter in the Al-Quran. To help the readers of the Al- Quran in pondering the meaning of the Al-Quran and linking the verses to the meaning of the word or the interpretation of the verse in a tafsir book, we propose the first multi-layer Quranic Knowledge Graph that uses the property graph format. We also add a chatbot on top of our Quranic Knowledge Graph. And to save cost in using a Large Language Model, we deploy a vector database to memorize the previously answered user queries and their corresponding Cypher translations. We evaluate our approach using three point of views: large language model, knowledge graph, and chatbot. The result of evaluation shows that our Quranic Knowledge Graph achieves 100% correctness and 90% accuracy, despite only covering 9 parts of the Al-Quran. For the interested readers, please access https://zentilax.github.io/quranic-chatbot-UI/ for exploring the quranic knowledge graph