ABSTRACT
Detection of cyberbullying that often occurs on social networks, for instance, twitter, is one of the targeted areas of research nowadays. Due to the easy accessibility and popularity of social media, cyberbullying has become a major challenge in electronic communication, while the behavior of cyberbullying has received increasing attention recently. A set of input data is sent to the data processing unit that is incorporated to improve the quality of input data. Data preprocessing also involves the removal of keywords and special characters. After processing the data, the output data is sent to feature extraction phase. This research work leveraged feature ranking method to select the options from tweets. The advanced deep learning model of oppositional grasshopper optimization with convolutional neural network (OGHOCNN) was used in the next step to detect lucky messages for cybercrime. Performance measurements, for instance, classification accuracy, were selected to assess the effectiveness of OGHOCNN classifiers. The model was implemented in MATLAB tool and the results exhibited promising outcomes.
