Abstract
The rise of social networking in today’s society has brought convenience to people’s lives, but at the same time people are also suffering from cyberbullying. How to check these bullying languages has become a popular problem. As text is an important vehicle for online social networking, the natural language learning, representation, and training becomes a necessary work for cyberbullying detection. In this paper, we summarize and analyze the existing work by studying it, and then finally propose new ideas and experiments. The specific method is based on the LSTM model, in which the parameters and dimensions are adjusted to demonstrate the best results of the model. And a user rating system is used to detect bullying more effectively.