As cyber threats increasingly propagate through Arabic-language online platforms, the need for effective Arabic natural language processing tools for cybersecurity applications has become critical. This paper presents a deep learning approach to cybersecurity threat detection in Arabic content, addressing the unique linguistic challenges of Arabic text processing including morphological richness, dialectal variation, and right-to-left script handling. We evaluate multiple deep learning architectures including convolutional neural networks, recurrent neural networks, and transformer-based models for Arabic threat classification tasks. The proposed approach integrates Arabic-specific preprocessing techniques with contextual word embeddings to capture the semantic nuances of threat-related content. Experimental evaluation on a curated dataset of Arabic cybersecurity content demonstrates that the transformer-based architecture achieves classification accuracy of 93.4% for threat detection, outperforming traditional machine learning baselines by 18.7 percentage points. We also present a comparative analysis of different Arabic text representation strategies and their impact on cybersecurity classification performance.
Mohamed et al. (Sun,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: