KurdSum: A new benchmark dataset for the Kurdish text summarization
Authors
Natural Language Processing Journal
Vol. 5
pp. 100043
December 1, 2023
Abstract
Introduces KurdSum, a Kurdish news summarization benchmark containing more than 40,000 texts paired with human-written summaries. The study evaluates four extractive approaches (LexRank, TextRank, Oracle, and Lead-3) and three abstractive approaches (Pointer-Generator, Sequence-to-Sequence, and Transformer-based summarization), with Pointer-Generator achieving the strongest abstractive ROUGE results and Oracle the strongest extractive results.
Keywords
Kurdish
Automatic Text Summarization
KurdSum
News
ROUGE
Low-Resource NLP
Related Datasets
KurdSum Dataset
Date not verified
40,000+ article-summary pairs
Text, Article-summary pairs
A Kurdish news summarization benchmark with more than 40,000 news texts paired with human-written summaries. It was introduced with the KurdSum study and is publicly archived in Mendeley Data.
Citation
Badawi, S. (2023). KurdSum: A new benchmark dataset for the Kurdish text summarization. Natural Language Processing Journal, 5, 100043.