Toxicity and Sentiment Analysis About Digital Bounty on Social Media

Abstract

The presence of social media has a positive impact but social media is also a means for negative things to happen, such as cyberbullying, data theft, and sexual crimes. Digital bounty or competition to find perpetrators digitally is one of the activities of cyberbullying and digital vigilantism (doxing). Digital bounties have negative impacts such as psychological and socio-economic impacts on perpetrators (bounty targets) and encourage reduced public trust in applicable legal institutions. The number of opinions on social media requires sentiment classification and toxicity measurements. To get the best results, a classification method is searched for the appropriate dataset by comparison, using k-fold validation (KNN, NBC, SVM, DT), in the classification process using the selected method. Give positive opinion results, (supporting digital bounty posts of 1540 and negative of 439, and in toxicity, analysis using Communalytic, most comments in the dataset scored 0 – 0.4