Multi-Label Theme-Based Classification of Quranic Verses for Urdu TranslationsA Case Study for Surah Al-Baqarah and Aal-e-Imran

Abstract

The Holy Quran stands as our timeless source of guidance, which explores multiple deep subjects for understanding all aspects of human life and moral behavior. The research focuses on the Quranic verses from the Urdu translation of Surah Al-Baqarah and Aal-e-Imran into different themes to help with better topic analysis. The research aims to develop a thematic classification system for Quranic verses by utilizing machine learning techniques and evaluating the system's performance through standard classification metrics. Using XLM-Roberta Classifier is the powerful tool model to translate in Urdu language by Junagharri from the Tanzil website, it translates it with standard measure of performance. This model give produces the 0.9501 F1 score along with 0.8936 accuracy, 0.9726 precision, and 0.9286 recall rates. The result shows the performance and excellent working when they handling multiple labels for analyzing religious texts. This research provides a detail knowledge to scholars and students of the Quranic verses by different useful tools and open the doors for the progress and progress in using AI for religious studies.