EUVIMEDEUVIMED entdecken
Crossref

Hierarchical Classification of Arabic Media: Employing a Two-Task BERT Model

Aisha Blfgeh, Hanen Himdi, Rehab Qarout, Fatmah Assiri, Amjad Qashlan

International Journal of Advanced Computer Science and Applications · 2026 · 17 (10)

DOI: 10.14569/ijacsa.2026.01710107 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Advanced Computer Science and Applications, 2026. Verfasst von Aisha Blfgeh und 4 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Thousands of news articles are published on the web daily, causing the need to filter and extract knowledge on specific topics. With the increasing proliferation of digital content, automatically differentiating between informative news reporting and opinion-based columns has become essential to support fact-checking, content moderation, and journalistic integrity. Although previous studies have focused mainly on classifying news articles by domain in languages such as English, this research advances Arabic news classification by proposing a novel hierarchical two-task model to distinguish between news articles and columns. The proposed model leverages AraBERT, a pretrained Arabic transformer-based language model, to first differentiate between news and columns, followed by the classification of news articles into six categories: Economy, Health, Local, Politics, Sports, and Technology. To support model development, a comprehensive dataset comprising 7,367 columns was collected from various Arabic news platforms. Experimental results demonstrate that the model achieved 95% accuracy in distinguishing news from columns and 94% accuracy in classifying news articles into their respective domains. When tested on real-world articles, the model attained 96% accuracy in the first classification task and 75% accuracy in domain classification. Additionally, an in-depth performance analysis was conducted at both the class and word levels using feature importance and attention mechanisms, respectively, to better understand the linguistic characteristics that differentiate news from columns. These findings underscore the model’s effectiveness and highlight its potential for future optimizations, particularly in enhancing Arabic text classification and improving the nuanced differentiation between journalistic writing styles.

Zitieren

APA

Blfgeh, A., Himdi, H., Qarout, R., Assiri, F., & Qashlan, A. (2026). Hierarchical Classification of Arabic Media: Employing a Two-Task BERT Model. International Journal of Advanced Computer Science and Applications, 17(10). https://doi.org/10.14569/ijacsa.2026.01710107

Vancouver

Blfgeh A, Himdi H, Qarout R, Assiri F, Qashlan A. Hierarchical Classification of Arabic Media: Employing a Two-Task BERT Model. International Journal of Advanced Computer Science and Applications. 2026;17(10). doi:10.14569/ijacsa.2026.01710107

BibTeX
@article{Blfgeh2026,
  author = {Blfgeh, Aisha and Himdi, Hanen and Qarout, Rehab and Assiri, Fatmah and Qashlan, Amjad},
  title = {Hierarchical Classification of Arabic Media: Employing a Two-Task BERT Model},
  journal = {International Journal of Advanced Computer Science and Applications},
  year = {2026},
  volume = {17},
  number = {10},
  doi = {10.14569/ijacsa.2026.01710107}
}
Kostenlos starten →