Analisis Sentimen Twitter Terhadap Cyberbullying Menggunakan Metode Support Vector Machine (SVM)

Abstract

The development of technology helps humans to communicate with each other. The rapid development of technology has led to the birth of many new media platforms. As technology advances in this digital era. The existence of information technology has a great impact on human civilization. One form is the use of social media. However, the ease of sharing information through social media has not escaped abuse by its users. One form of abuse is cyberbullying carried out on social media, especially Twitter. The 2018 Internet Penetration and Internet User Behavior Survey in Indonesia published by the Indonesian Internet Service Providers Association (APJII) shows that 49 percent of internet users have experienced bullying in the form of ridicule or harassment on social media. In this regard, sentiment analysis was carried out through social media twitter to measure the accuracy value using the SVM (support vector machine) algorithm. Sentiment analysis by crawling twitter data as many as 1000 tweet data in Indonesian using the RapidMiner tool. And this classification with SVM (Support Vector Machine) algorithm gets an accuracy of 92% by testing at 80:20 proportion, which is 80% training data and 20% testing data. Of the 998 data that have been preprocessed, 625 data on the positive class and 374 data on the negative class have been obtained.