Deep Learning Cyberbullying Detection Using Stacked Embbedings Approach

The Cyberspace is one of the humanity's great inventions that bring great benefits but also exposes us to cyber threats. Cyberbullying commonly happened to each and every person on social platforms. In this paper we propose a framework to detect cyberbullying messages in the form of text data using deep neural networks and word embeddings. We stack together the state-of-the-art Bert and Glove embeddings to improve the performance of the classifier. As a result, the model outperforms the majority of the traditional machine learning methods such as SVM and Logistic Regression.