An efficient cyberbullying detection framework on social media platforms using a hybrid deep learning model

People in social media are more vulnerable to the negative effects and the most serious consequences of utilising social media is cyberbullying. Cyberbullying is an offensive and deliberate act perpetrated online by a particular individual or an organisational structure. It is brought about by sending, publishing, and disseminating offensive, dangerous, and misleading information online. As cyberbullying becomes increasingly prevalent in social media, automatically detecting it and taking proactive steps to address it becomes critical. Humiliation of an individual in social media causes psychological disturbance in one's life, in order to have a safe and secure platform. A hybrid deep learning model has been used that combines convolutional neural network (CNN) and long short-term memory (LSTM) to detect cyberbullying more precisely and effectively in this paper. Using convolutional layers and max-pooling layers, the CNN model recovers higher level features efficiently. Long-term dependencies between word sequences can be captured using the LSTM model. The findings reveal that in terms of accuracy, the presented hybrid CNN-LSTM Model performs better than standard approaches for machine learning and deep learning.