Now in modern culture, the presence of social media takes communication to a whole new level with social networking. People can speak freely on social media without much concern. However, this also causes users to be not aware of the negative effects of cyber-aggression towards victim in long term. Users may not be aware that their posts or comments on Facebook contain offensive messages to others. There are many cyber-aggression cases appearing on current social media platforms because they do not provide cyberbully detection features or prevention mechanisms to avoid harassment incidents happen on social media. Cyberbullies are free to use any social media platform to bully other users without much restriction, even the current existing cyber-bullying detection systems also hard to provide significant impact on detecting aggressive behaviours happened online as most of them did not provide data visualization and have many limitations on text processing and cyber- aggression detection. These reasons contribute this project proposed to develop the application that use better machine learning models on cyber-aggression detection while providing more features than current existing detection systems.

Cyber-aggression Detection Application - FaceCyber is a combination of Google Chrome web extension and web application that use machine learning to detect cyber-aggression on Facebook. This application can contribute to reduce the impact of cyber-aggression on the Facebook which will be significant on the cyberbullies who did not have intention to hurt others online but lack of awareness on aggressive actions done by themselves. This detection system also can detect cyber-aggression effectively and stop online hate before the damage is done. Next, FaceCyber applies machine learning and natural language processing techniques to automatically extract offensive messages from text. One of the advantages of machine learning model is the model can be reused for further investigation and development after this project. The cyber-aggression detection model can be trained and reused on different social media applications, then Facebook will not the only application that supported by the cyber-aggression detection system.