Detection of Depression Among Arabic Twitter Users Using a Convolutional Neural Network
Proceedings of Engineering and Technology Innovation · 05.08.2026 · 34 · 110-122
DOI: 10.46604/peti.2026.16024 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of Engineering and Technology Innovation, online erschienen am 05.08.2026. Verfasst von Mohammed Majid Msallam und 2 weiteren Autor:innen. Die Arbeit nennt 30 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).
Themen: Depression
Abstract (Originalsprache)
Depression is a prevalent mental health condition, where early detection using the analysis of patients' feelings contributes to preventing and treating the exacerbation of symptoms. This work aims to detect depression in Arabic tweets using convolutional neural networks (CNN) trained on a dataset generated from the Arabic_Dep_tweets_10,000 dataset. Two datasets are generated using the stripping and captioning process to improve the generalization of the proposed CNN model. The stripping process removes punctuation marks from Arabic tweets, while the captioning process types the Arabic tweet on a white background to generate an image. Experimental results demonstrate that the proposed CNN model achieves up to 100% accuracy on the generated dataset treated with stripping and the captioning process.
Zitieren
APA
Msallam, M. M., Tarkan, E. Y., & Salim, S. K. (2026). Detection of Depression Among Arabic Twitter Users Using a Convolutional Neural Network. Proceedings of Engineering and Technology Innovation, 34, 110-122. https://doi.org/10.46604/peti.2026.16024
Vancouver
Msallam MM, Tarkan EY, Salim SK. Detection of Depression Among Arabic Twitter Users Using a Convolutional Neural Network. Proceedings of Engineering and Technology Innovation. 2026;34:110-122. doi:10.46604/peti.2026.16024
BibTeX
@article{Msallam2026,
author = {Msallam, Mohammed Majid and Tarkan, Esraa Yahia and Salim, Sarah Kareem},
title = {Detection of Depression Among Arabic Twitter Users Using a Convolutional Neural Network},
journal = {Proceedings of Engineering and Technology Innovation},
year = {2026},
volume = {34},
pages = {110--122},
doi = {10.46604/peti.2026.16024}
}
EUVIMED