Sentiment Analysis of Twitter Discussions on Rafael Alun: Multinomial Naïve Bayes and Decision Tree Approach
Abstract
Political events in Indonesia often draw significant attention on the social media platform Twitter, with netizens using the medium to express their opinions and feelings. One such viral topic on Twitter revolves around Rafael Alun, a former Director General of Taxation at the Indonesian Ministry of Finance who has been implicated in a possible gratification scandal and subsequently investigated by the Corruption Eradication Commission (KPK). As a result, Rafael Alun's name has been trending on Twitter. By using sentiment analysis, it is possible to identify the prevailing sentiment elements within tweets related to Rafael Alun. The application of Multinomial Naïve Bayes and Decision Tree algorithms serves to determine the accuracy of sentiment analysis results derived from Twitter tweets data. The research process involves steps such as data pre-processing, data processing, classification and evaluation. The sentiment analysis of tweets containing the mention of "Rafael Alun" shows that the majority of the tweets express a negative sentiment. The accuracy rates obtained by implementing the Multinomial Naïve Bayes and Decision Tree algorithms are 77% and 72% respectively. It is worth noting that these percentages are relatively moderate due to the unbalanced distribution of positive, negative and neutral sentiments on the topic.