ABPC-NET: A CAPSULE-GUIDED HYBRID FRAMEWORK FOR ROBUST ARABIC-TEXT CLASSIFICATION

(Received: 7-Mar.-2026, Revised: 18-Apr.-2026, 10-May-2026 and 16-May-2026 , Accepted: 29-May-2026)

Authors Baqer M. Merzah, Jafar Razmara,

Keywords #Arabic text classification #Deep learning #Transformer models #Capsule networks #Natural-language processing (NLP)

Abstract Arabic Text Classification (ATC) remains challenging due to the Arabic language's morphological richness and semantic complexity. This paper proposes ABPC-Net, a hybrid framework integrating a frozen Arabic Transformer encoder, a Bidirectional LSTM, parallel multi-scale CNN branches and a lightweight capsule- inspired vector projection head for hierarchical feature integration. Evaluated on the SANAD dataset and its subsets (AlArabiya, AlKhaleej and Akhbarona) over five independent runs, ABPC-Net achieves mean accuracies of 97.00±0.04%, 99.14±0.10%, 98.40±0.10% and 95.59±0.12%, respectively. Under identical experimental conditions, the proposed framework consistently outperforms re-implemented frozen and fully fine-tuned AraBERT and MARBERT baselines. Cross-dataset evaluation on BBC Arabic and CNN Arabic further provides evidence of intra-domain transferability and rapid few-shot adaptability across Arabic news sources. The reported results are scoped to Modern Standard Arabic news classification.

References

[1] A. Alrayzah, F. Alsolami and M. Saleh, "AraFastQA: A Transformer Model for Question-answering for Arabic Language Using Few-shot Learning," Computer Speech & Language, vol. 95, p. 101857, 2026.

[2] A. Wali et al., "Evaluating Arabic Sentiment Analysis with GPT-4o: A Comparative Study of Raw and Pre-processed Text," J. of Cases on Inf. Tech., vol. 28, no. 1, pp. 1-19, 2026.

[3] J. H. Yousif, "Artificial Intelligence and Machine Learning for Enhancing Arabic Fake News Detection: A BERT-based Transformer Approach," Procedia Computer Science, vol. 275, pp. 809-816, 2026.

[4] E. Alnagi, R. Ghnemat and Q. Abu Al-Haija, "Boosting Arabic Text Classification Using Hybrid Deep Learning Approach," Discover Applied Sciences, vol. 7, no. 6, p. 540, May 2025.

[5] A. Elnagar, R. Al-Debsi and O. Einea, "Arabic Text Classification Using Deep Learning Models," Information Processing & Management, vol. 57, no. 1, p. 102121, Jan. 2020.

[6] N. Boudad, R. Faizi, R. Oulad Haj Thami and R. Chiheb, "Sentiment Analysis in Arabic: A Review of the Literature," Ain Shams Engineering Journal, vol. 9, no. 4, pp. 2479-2490, Dec. 2018.

[7] B. M. Merzah, J. Razmara and Z. Salmanian, "Hybrid Deep Learning Models for Fake News Detection: Case Study on Arabic and English Languages," Frontiers in Big Data, vol. 8, DOI: 10.3389/fdata.2025.1683786, Jan. 2026.

[8] A. B. Nassif et al., "Arabic Fake News Detection Based on Deep Contextualized Embedding Models," Neural Computing & Applications, vol. 34, no. 18, pp. 16019-16032, Sep. 2022.

[9] I. Guellil, H. Saâdane, F. Azouaou, B. Gueni and D. Nouvel, "Arabic Natural Language Processing: An Overview," J. of King Saud Uni.-Computer and Inf. Sciences, vol. 33, no. 5, pp. 497-507, 2021.

[10] A. Y. Muaad et al., "A Novel Deep Learning ArCAR System for Arabic Text Recognition with Character-level Representation," Computer Sciences and Mathematics Forum, vol. 2, no. 1, Article no. 14, DOI: 10.3390/IOCA2021-10903, and Presented at the 1st Int. Electronic Conf. on Algorithms, Sep. 2021.

[11] M. Louail et al., "Tasneef: A Fast and Effective Hybrid Representation Approach for Arabic Text Classification," IEEE Access, vol. 12, pp. 120804-120826, 2024.

[12] I. Jamaleddyn, R. ayachi and M. Biniz, "Novel Multichannel Deep Learning Model for Arabic News Classification," Jordanian Journal of Computers and Information Technology (JJCIT), vol. 10, no. 4, pp. 453-468, DOI: 10.5455/jjcit.711720086134, Dec. 2024.

[13] Md. M. Hossain et al., "A Hybrid Attention-based Transformer Model for Arabic News Classification Using Text Embedding and Deep Learning," IEEE Access, vol. 12, pp. 198046-198066, 2024.

[14] M. El Kourdi, A. Bensaid and T. E. Rachidi, "Automatic Arabic Document Categorization Based on the Naïve Bayes Algorithm," Proc. of the Workshop on Computational Approaches to Arabic Script-based Languages (Semitic '04), pp. 51-58, Geneva, Switzerland, 2004.

[15] L. Al Qadi, H. El Rifai, S. Obaid and A. Elnagar, "A Scalable Shallow Learning Approach for Tagging Arabic News Articles," Jordanian Journal of Computers and Information Technology (JJCIT), vol. 6, no. 3, pp. 263-280, DOI: 10.5455/jjcit.71-1585409230, 2020.

[16] A. Y. Muaad et al., "An Effective Approach for Arabic Document Classification Using Machine Learning," Global Transitions Proceedings, vol. 3, no. 1, pp. 267-271, Jun. 2022.

[17] D. Alsaleh and S. Larabi-Marie-Sainte, "Arabic Text Classification Using Convolutional Neural Network and Genetic Algorithms," IEEE Access, vol. 9, pp. 91670-91685, 2021.

[18] M. S. H. Ameur, R. Belkebir and A. Guessoum, "Robust Arabic Text Categorization by Combining Convolutional and Recurrent Neural Networks," ACM Trans. on Asian and Low-resource Language Information Processing, vol. 19, no. 5, pp. 1-16, DOI: 10.1145/3390092, Sep. 2020.

[19] A. A. Jalil and A. H. Aliwy, "Classification of Arabic Social Media Texts Based on a Deep Learning Multi Tasks Model," Al-Bahir, vol. 2, no. 2, DOI: 10.55810/2313-0083.1030, May 2023.

[20] B. B. Al-onazi et al., "Automated Arabic Text Classification Using Hyper-parameter Tuned Hybrid Deep Learning Model," Computers, Materials & Continua, vol. 74, no. 3, pp. 5447-5465, 2023.

[21] W. Antoun, F. Baly and H. Hajj, "AraBERT: Transformer-based Model for Arabic Language Understanding," Proc. of the 4th Workshop on Open-Source Arabic Corpora and Processing Tools, with a Shared Task on Offensive Language Detection, pp. 9-15, Marseille, France, 2020.

[22] A. Jalili, H. Tabrizchi, J. Razmara and A. Mosavi, "BiLSTM for Resume Classification," Proc. of the 2024 IEEE 22nd World Symposium on Applied Machine Intelligence and Informatics (SAMI), pp. 519-524, Stará Lesná, Slovakia, 2024.

[23] Y. Kim, "Convolutional Neural Networks for Sentence Classification," Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), pp. 1746–1751, Doha, Qatar, DOI: 10.3115/v1/D14-1181, 2014.

[24] S. Sabour, N. Frosst and G. E. Hinton, "Dynamic Routing between Capsules," Proc. of the 31st Conf. on Neural Information Processing Systems (NIPS 2017), Long Beach, CA, USA., vol. 30, 2017.

[25] D. M. W. Powers, "Evaluation: From Precision, Recall and F-measure to ROC, Informedness, Markedness and Correlation," arXiv: 2010.16061, 2020.

[26] M. Benhammouda, A. Khobzaoui and N. Mahammed, "Arabic Text Classification Using Graphs and Deep Learning," IJCESEN, vol. 11, no. 4, pp. 9415-9421, DOI: 10.22399/ijcesen. 4402, 2025.

[27] M. Berrimi, M. Oussalah, A. Moussaoui and M. Saidi, "A Comparative Study of Effective Approaches for Arabic Text Classification," Available at SSRN 4361591, DOI: 10.2139/ssrn. 4361591, Feb. 2023.

[28] M. S. A. Alzaidi et al., "Enhanced Automated Text Categorization via Aquila Optimizer with Deep Learning for Arabic News Articles," Ain Shams Engineering Journal, vol. 16, no. 1, p. 103189, DOI: 10.1016/j.asej.2024.103189, Jan. 2025.

[29] M. K. Saad and W. Ashour, "OSAC: Open Source Arabic Corpora," Proc. of the 6th ArchEng Int. Symposiums on Electrical and Electronics Engineering and Computer Science (EEECS’10), vol. 10, p. 55, DOI: 10.13140/2.1.4664.9288, 2010.