Holy Quran Kurdish Sorani translation dataset for language modelling

Area of Science:

• Computational Linguistics
• Islamic Studies
• Religious Text Analysis

Background:

• The Holy Quran is a foundational Islamic text with global translations.
• Existing Kurdish translations may not meet the specific needs of NLP and linguistic analysis.
• A high-quality, semantically accurate translation is crucial for computational research.

Purpose of the Study:

• To introduce a manually translated Kurdish version of the Holy Quran.
• To create a dataset suitable for natural language processing (NLP) research.
• To facilitate linguistic analysis and computational studies on under-resourced languages.

Main Methods:

• A multi-year translation process involving bilingual scholars, translators, and proofreaders.
• Integration of advanced linguistic tools to ensure semantic accuracy and theological precision.
• Development of a dataset comprising raw and refined linguistic versions of the translation.

Main Results:

• A comprehensive Kurdish Quran translation dataset is now available.
• Statistical analyses included word frequency, verse length comparison (Kurdish vs. Arabic), and unique word distribution.
• The dataset demonstrates high fidelity to the original Arabic text's meaning and theological context.

Conclusions:

• This Kurdish Quran translation dataset is a valuable resource for computational linguistics.
• It supports the development of neural machine translation models for Kurdish.
• The dataset advances linguistic research, particularly for under-resourced languages like Kurdish.