University of Zakho Researchers Publish the First Open Badini Kurdish Corpus
A 1.28-million-word open corpus, paired with a purpose-built preprocessing toolkit, lays the foundation for Kurdish natural language processing and artificial intelligence — with transformative potential for the Badini dialect.
The Semantic Web Lab (SWLab) at the University of Zakho has published a landmark study in Data in Brief, an international peer-reviewed journal by Elsevier, one of the world’s leading scientific publishers. Titled “Dasinya: A Dataset and Preprocessing Pipeline for the Badini Kurdish Dialect,” the work delivers the first openly available Badini Kurdish text corpus — more than 1.28 million words across 87,545 sentences and 107 documents, spanning nine genres from literature and history to news and academic writing.
What sets this release apart is that the corpus does not arrive alone. It is accompanied by a complete, purpose-built preprocessing toolkit engineered specifically for Badini — capable of resolving the orthographic and Unicode challenges that general Arabic, Persian, and Sorani tools cannot. Both are fully open and free for anyone to use, giving researchers, developers, and institutions worldwide the resources to build and strengthen Kurdish-language technology.
The significance runs deep. This work forms the groundwork for Kurdish Natural Language Processing (NLP) — the essential layer on which artificial intelligence for the Kurdish language depends. It serves both the Sorani and Badini dialects and stands to be genuinely transformative for Badini Kurdish, which until now has had almost no digital presence. From tokenizers and part-of-speech taggers to large language models and AI systems, the next generation of Kurdish-language tools can now be built on open, reproducible resources.
This publication is the newest achievement in a continuous research initiative launched in 2020 to bring the Kurdish language — both Sorani and Badini in the Perso-Arabic (Arabic) script — into modern technology. Over the past several years the initiative has produced a steady stream of peer-reviewed publications, with this article marking its latest milestone.
The Directorate warmly congratulates the research team — Ms. Vaman A. Saeed and Dr. Karwan Jacksi — and extends its sincere appreciation to the Zakho Center for Kurdish Studies, directed by Prof. Dr. Abdulsalam and part of the General Directorate of Scientific Research, for its valued collaboration.
The Semantic Web Lab operates under the General Directorate of Scientific Research at the University of Zakho, directed by Prof. Dr. Dindar Bari.
Access the research and resources:
- Paper: https://www.sciencedirect.com/science/article/pii/S2352340926006761
- Open dataset & toolkit (Zenodo): https://doi.org/10.5281/zenodo.20729060
- Semantic Web Lab: https://swlab.uoz.edu.krd/
- General Directorate of Scientific Research: https://web.uoz.edu.krd/centers/dsrc/