KurdishMCQ: A multiple-choice question dataset for the Kurdish language (Sorani dialect)
Data in Brief · 2026-11 · 113409
DOI: 10.1016/j.dib.2026.113409 ↗Kurz gefasst
Fachartikel in der Zeitschrift Data in Brief, 2026-11. Verfasst von Khalid H. Ahmed und 1 weiteren Autor:innen. Die Arbeit nennt 16 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
This article presents a large-scale multiple-choice question (MCQ) dataset for the Kurdish language (Sorani dialect), designed to support research in natural language processing (NLP) and low-resource language modeling. The dataset was constructed from authentic educational materials, including middle and high school textbooks, end-of-chapter exercises, and official examination questions prepared by educational authorities. As these resources were not available in digital format, the questions were digitized using an optical character recognition (OCR) pipeline specialized for Sorani Kurdish and verified by native Kurdish speakers. Each instance includes the question text, multiple-choice options, the correct answer, and metadata such as subject and grade level. Only text-based questions were included; items requiring external context such as images or tables were excluded. The dataset comprises 17,221 questions spanning biology, chemistry, mathematics, physics, natural science, history, geography, economics, Kurdish grammar and literature, and religion studies. The questions primarily correspond to grades 9–12, with additional categories such as driving tests. A multi-stage quality-control process was applied, including duplicate removal, manual review, and external validation by native speakers. The dataset can be used for multiple NLP tasks, including question answering, knowledge probing, reasoning evaluation, and multilingual model assessment, and is suitable for training, fine-tuning, and benchmarking large language models, particularly in zero-shot and few-shot settings. By offering a comprehensive and culturally grounded collection of Kurdish-language questions, this resource contributes to expanding data availability for underrepresented languages and facilitates research in multilingual and low-resource NLP.
Zitieren
APA
Ahmed, K. H., & Rashid, T. A. (2026). KurdishMCQ: A multiple-choice question dataset for the Kurdish language (Sorani dialect). Data in Brief, 113409. https://doi.org/10.1016/j.dib.2026.113409
Vancouver
Ahmed KH, Rashid TA. KurdishMCQ: A multiple-choice question dataset for the Kurdish language (Sorani dialect). Data in Brief. 2026:113409. doi:10.1016/j.dib.2026.113409
BibTeX
@article{Ahmed2026,
author = {Ahmed, Khalid H. and Rashid, Tarik A.},
title = {KurdishMCQ: A multiple-choice question dataset for the Kurdish language (Sorani dialect)},
journal = {Data in Brief},
year = {2026},
pages = {113409},
doi = {10.1016/j.dib.2026.113409}
}
EUVIMED