A GRU-Driven Machine Learning Model for Detecting Cyberbullying on Twitter

One of the biggest issues facing the safe online community is the exposure of cyberbullying on social media, especially on Twitter. This work explores in depth how the Gated Recurrent Unit(GRU) method can be used to classify and identify cyberbullying content in Twitter datasets. The GRU algorithm can be used in an application to determine whether the given status updates can be classified as bullying or non-bullying with the help of patterns and rules. It is evaluated by testing its performance on the accuracy of risk information. The results show that the GRU algorithm performs well in automatically detecting cyberbullying cases, which is useful for the development of more robust social media management systems.