Hyperparameter Tuning of Pre-Trained Architectures for Multi-Modal Cyberbullying Detection
Subbaraju Pericherla, Lakshmi Hyma Rudraraju, Nirbhay Kumar Chaubey · Advances in information security, privacy, and ethics book series · 2024
Over the past decade, cyberbullying has become a pervasive issue, particularly among young individuals, causing growing concern within society. The rise of social media provided fertile ground for cyberbullying incidents to occur. In this work, proposed a deep learning based Multi-Modal Cyberbullying Detection(MMC) technique to identify cyberbullying on both text and image data combination. This MMC technique involves two pre trained deep architectures for generate feature vector representations. The RoBERTa and Xception architectures are employed to extract features from the text data and the image respectively. LightGBM classifier is used to classify the multi-modal data is bullying or non-bullying . The hyperparameter tuning is applied RoBERTa and Xception architectures to improve classification performance of MMC for cyberbullying detection on multi-modal data. The experiments conducted on 2100 samples of combined data of text and image. The proposed MMC technique efficiently classifies bullying data with f1-score of 80% and outperforms as compared to existing approaches.