Arabic Fake News Detection - Using Machine Learning Approach

Aug 3, 2026

Uploaded by Noah

Widespread dissemination of false Arabic news creates serious social risks that damage digital information credibility and erode trust. This study applies machine learning to distinguish authentic Arabic news from deceptive counterparts, emphasizing how misinformation threatens social cohesion and the foundations of societal bonds. The work addresses Arabic text classification challenges caused by complex morphology and many diacritical marks by preprocessing two datasets for training and testing. Models include Random Forest, Logistic Regression, Decision Tree, and Multinomial Naive Bayes, with Logistic Regression achieving 95.92% accuracy on the SANAD dataset.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages8

Reading Time20 min

AccessFree to download

Abstract

Introduction

Methodology and Datasets

Machine Learning Models and Results

Please cite the Published Version Duridi, Tasneem, Eleyan, Derar, Eleyan, Amna and Bejaoui, Tarek (2024) Arabic Fake News Detection Using Machine Learning Approach. In: 2024 International Symposium on Networks, Computers and Communications (ISNCC), 22 October 2024-25 October 2024, Washington DC, USA. DOI: [https://doi.org/10.1109/isncc62547.2024.10758936](https://doi.org/10.1109/isncc62547.2024.10758936) Publisher: IEEE Version: Accepted Version Downloaded from: [https://e-space.mmu.ac.uk/637496/](https://e-space.mmu.ac.uk/637496/) Usage rights: In Copyright Additional Information: 2024 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. Enquiries: If you have questions about this document, contact [openresearch@mmu.ac.uk. Please](openresearch@mmu.ac.uk. Please) include the URL of the record in e-space. If you believe that your, or a third party's rights have been compromised through this document please see our Take Down policy (available from [https://www.mmu.ac.uk/library/using-the-library/policies-and-guidelines](https://www.mmu.ac.uk/library/using-the-library/policies-and-guidelines)) Arabic Fake News Detection using Machine Learning Approach 1st Tasneem Duridi Department of Computer Science Palestine Technical University Tulkarm, Palestine [tasneem.duridi@ptuk.edu.ps](tasneem.duridi@ptuk.edu.ps) 2nd Derar Eleyan Faculty of Telecomm. and Information Technology Nablus University for Technical and Vocational Education Nablus, Palestine [d.eleyan@nu-vte.edu.ps](d.eleyan@nu-vte.edu.ps)[ ](d.eleyan@nu-vte.edu.ps)Department of Computer Science Palestine Technical University -Kadoorie Tulkarm, Palestine [d.eleyan@ptuk.edu.ps](d.eleyan@ptuk.edu.ps) 3rd Amna Eleyan Department of Computing and Mathematics Manchester Metropolitan University Manchester M1 5GD, UK[a.eleyan@mmu.ac.uk](a.eleyan@mmu.ac.uk) 4th Tarek Bejaoui Computer Engineering Department University of Carthage Tunisia [tarek.bejaoui@ieee.org](tarek.bejaoui@ieee.org) Abstract—In the contemporary digital realm, the widespread dissemination of false Arabic news is a significant social concern laden with various risks. Recognizing the seriousness of this issue, our research utilizes cutting-edge technologies, specifically Machine Learning, to distinguish between authentic Arabic news and deceptive counterparts. The consequences of propagating misinformation go beyond compromising social cohesion; they extend to the erosion of digital information’s credibility, fostering an atmosphere of mistrust and deception that undermines the very foundations of societal bonds. Through the application of contemporary technologies, this study aims to identify and underscore Arabic news that embodies such risks. The intricate characteristics of Arabic language morphology, marked by words carrying multiple meanings based on inflectional forms and the prevalence of numerous diacritical marks, intensify the challenges of text classification. D espite contending with these linguistic intricacies, modern natural language processing approaches offer practical solutions. Notably, our methodology relies on the preprocessing of two available datasets for training and testing, a crucial step for seamlessly integrating a range of Machine Learning techniques, including Random Forest (RF), Logistic Regression (LR), Decision Tree (DT), and Multinomial Naive Bayes (NB). Importantly, the Logistic Regression technique emerged as the most effective, achieving an accuracy of 95.92% in the SANAD dataset for discerning nuanced Arabic news Index Terms—Machine Learning, Arabic, Fake News, NLP, News, Fake News Detection, Text Classification, SANAD,

FAQs

Why is Arabic fake news detection considered a significant social concern?

False Arabic news dissemination increases misinformation risks, undermines social cohesion, and erodes trust in digital information. It also fosters an environment of mistrust and deception.

What challenges make Arabic text classification difficult in this study?

Arabic morphology introduces multiple meanings depending on inflectional forms, and the text includes numerous diacritical marks. These linguistic properties intensify classification complexity.

Which machine learning method performed best and what accuracy was reported?

Logistic Regression performed the best, reaching 95.92% accuracy on the SANAD dataset for discerning nuanced Arabic news.

Aug 3, 2026

Uploaded by Noah