Association Analysis of Cyberbullying on Social Media using Apriori Algorithm
Zuraini Zainol, Sharyar Wani, Puteri N.E. Nohuddin, Wan M.U. Noormanshah, Syahaneim Marzukhi · International Journal of Engineering & Technology · 2018
With the phenomenal increase in use of Social Networking Service (SNS) and mobile technology, the consequences of cyberbullying have become an epidemic. More than 80% youth use cell phones making them extremely vulnerable to the abuse and one in three young people have been found victims of this problem. There are many different methods of detection cyberbullying behaviour patterns however rarely any focuses on analysis based on association especially in Malay language. Learning and detecting using association is a natural communication phenomenon that can help to identify abusive content from the hidden corpora, which often goes unnoticed. Association helps to identify trends, rules and patterns of the bullies and detects abusive content considering whole sets rather than focusing on single instances. The current work focuses on detection of cyberbullying instances by association analysis using the Apriori Algorithm. It mainly focuses on detecting bullying and aggressive behaviour on Twitter. Over 80 different patterns with high confidence levels were detected that can be successfully implemented for the detection process. The high confidence levels are indicative of the efficiency of association analysis for cyberbully detection in SNS.