Cyberbullying, hostile behavior of a group or an individual to defame or harass the victim mentally with the help of social media and other e-communication platforms, has the potential to create a lifelong negative impact on mental health with the power of inducing suicidal thoughts. It is on the rise among the early adolescents of the age group from 8 to 16. Hence it is vital to detect Cyberbullying at an early stage to safeguard the victims at the high risk of developing depression, anxiety, and suicidal ideas. It also helps to mitigate psychological, academic, and social consequences. Existing cyberbullying detection approaches primarily depend on static monolingual questionnaires and are not personalised. With the developments in Artificial Intelligence, many neural network-based approaches are explored to detect cyberbullying. This study discusses and provides comparative analysis of various machine learning approaches for detecting cyberbullying victimization among school students highlighting their effectiveness and limitations.