Classification of Cyberbullying in Facebook Using Selenium and SVM

Abstract

Cyberbullying is one of the emerging problems over the past few years especially to teenagers. Approximately 24% of teens goes online constantly, facilitated by the widespread availability of smartphones. Almost 21% of teens said the main reason they checked social media always was to make sure nobody was saying mean or bad things to them. Cyberbullying related Facebook posts were harvested by a customized web scraper tool. These harvested data were used for classification using Support Vector Machines (SVM) model. A total of 2263 data was used for training data, Facebook posts. Based on these posts, the study achieved the precision of 88% and the recall is 87%.