Hierarchical Arabic Text Classification: BERT-ANN vs. BERT-BiLSTM

Text classification, the task of assigning text to predefined categories, is a fundamental area in natural language processing, often requiring sophisticated techniques and systems. Hierarchical text classification, which structures categories in a tree-like hierarchy, introduces additional challenges due to the interplay of similarities and relationships between categories across different levels. The application of artificial intelligence (AI) techniques has significantly advanced text classification, but addressing this task in Arabic remains particularly challenging due to limited linguistic resources. In this paper, we propose and evaluate two AI-based models, BERT-ANN and BERT-BiLSTM, for hierarchical Arabic text classification. Experimental results reveal that the BERT-ANN-based model achieves superior performance on short text classification, obtaining an HF-score of 92%, compared to 77% achieved by the BERT-BiLSTM-based model.