Cyberbullying Detection Approaches: A Review

Bullying today is defined as repeated and intentional hostile actions by a powerful person against one more vulnerable. It frequently occurs online, due to the proliferation of digital devices and widespread internet use, particularly among young people. Social media refers to highly popular sharing platforms that obtain millions of views, such as Twitter. These receive several comments, containing different information. These comments are utilised using natural language processing techniques and machine learning techniques. Many approaches for the detection of cyberbullying have been proposed in the literature, including machine learning, deep learning and hybrid learning, which can be used to identify bullying phrases. However, it is challenging to identify the most accurate model. Attempts have also been made to detect cyberbullying in social comments by identifying the bullying words in Arabic languages as well. This research study investigates the existing methods and techniques from different approaches that can be used on social media content to classify cyberbullying. Additionally, it explains and discussion the approaches which are useful in text mining and classifying offensive words.