DIGITAL QUR’AN ANNOTATION THROUGH DEEP LEARNING: TOWARD AN INTELLIGENT TAFSIR ECOSYSTEM

Abstrak

Traditional Qur’anic exegesis (tafsir) encompasses vast, complex textual datasets. Current digital platforms provide static access but lack the semantic understanding to intelligently connect, compare, and synthesize diverse interpretative traditions, limiting deep analysis for both scholars and lay users. This study aims to develop and evaluate a deep learning framework for the semantic annotation of Qur’anic verses. We seek to automate the classification of verses based on thematic content, legal rulings (ahkam), and intertextual references to classical tafsir, thereby prototyping an intelligent tafsir ecosystem. A gold-standard annotated corpus was constructed by mapping Qur’anic verses to thematic categories from authoritative tafsir. A custom transformer-based (BERT) model was trained on this corpus to perform multi-label classification and evaluated using F1-score, precision, and recall. The model demonstrated high efficacy, achieving a macro F1-score of 0.92 in thematic annotation and 0.89 in identifying intertextual links. The system accurately predicts and suggests relevant tafsir passages for un-annotated verses, significantly outperforming traditional keyword-based search methods in relevance. This study validates deep learning as a tool for sophisticated digital Qur’an annotation. The framework provides the foundation for an intelligent tafsir ecosystem, capable of offering dynamic, contextual, and deeply networked access to Qur’anic knowledge.

Identifier

• DOI: https://doi.org/10.70177/ijonis.v2i5.2572
• OpenAlex: https://openalex.org/W7169801770