Hinglish Abusive Comment Detection Using Transformer-Based Models
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Hinglish (a combination of Hindi and English), is in prevalent use online in India. Detection and flagging of abusive content in a code switched language is complex due to transliteration, variability, and informal slang usage. In this study, a hybrid deep learning approach is proposed to deal with these challenges through the integration of transformer-based models, specifically Multilingual BERT (mBERT) and XLM-RoBERTa, with a bidirectional GRU (BiGRU) layer for enhancement of sequential processing. The attained F1 scores of the proposed models are: mBERT with a score of 0.8649 and XLM-RoBERT with score of 0.8664.
Fields of Study & Subjects
Hate Speech and Cyberbullying Detection
Sentiment Analysis and Opinion Mining
Topic Modeling
Publication details
- Journal / source
- International Conference on Advances in Intelligent Computing and Applications (AICAPS)
- Pages
- 1–6
- Publication date
- 11 February 2026
- Language
- EN