A comparative study of deep learning methods for cyberbullying detection

Tóm tắt

This paper conducts a comparative study of machine learning and deep learning approaches for cyberbullying detection on social networking platforms. The evaluated models include traditional classifiers such as Logistic Regression and Support Vector Machine (SVM), as well as deep learning architectures including LSTM, BiLSTM, CNN, and a hybrid CNN-BiLSTM model. Experimental results indicate that while SVM and Logistic Regression achieve competitive performance among traditional methods, the proposed CNN-BiLSTM model consistently outperforms others by effectively capturing both local and sequential text features. These findings demonstrate the effectiveness of integrating convolutional and recurrent neural networks in improving the accuracy and robustness of automated cyberbullying detection systems.