| dc.description.abstract | Along with the development of digital technology, the use of social media among the public is very easily accessible without limits. Through social media users ranging from individuals, groups, etc. can socialize and communicate. Social media provides freedom of expression to users, especially the freedom to comment on content that is seen and watched. However, the freedom to comment on social media is widely misused by irresponsible parties to mock, make fun of, spread hate speech, and provide negative statements containing elements of Religion, Race, Ethnicity, Physical and Sexual or what is often referred to as Cyberbullying. The existence of cyberbullying on social media raises concerns and negative impacts on social media users who are victims. The number of cyberbullying cases through comments with new variations of words, causes difficulty in recognizing cyberbullying. So, to combat cyberbullying on social media, it's crucial to develop a system that can identify remarks that exhibit cyberbullying characteristics. The goal of this project is to develop a website and Chrome extension that can identify cyberbullying remarks on social media material. The words will be represented numerically using the K-Nearest Neighbor method and TF-IDF. The data used in this study were 4574 Indonesian comments crawled from the TikTok application. Through the evaluation results in this study, the model is able to detect cyberbullying comments with an accuracy value of 79% and classification of cyberbullying with labels of religion, race, ethnicity, physique, and sex with an accuracy value of 91%. | en_US |
