SENTIMENT ANALYSIS DUE TO "MUDIK" PROHIBITED OF COVID-19 THROUGH TWITTER

Abstract

“Mudik” is a habit every year for the people of Indonesia to return to their hometowns before the Eid. The existence of the Corona Virus pandemic (COVID-19) hit all over the world, including Indonesia, resulting in a ban from the government to do Mudik. Social media such as Twitter is often used as an expression of some people in commenting on something like the ban on Mudik. Comments on Twitter that are often known as tweets can be used as material for sentiment analysis. However, it is not easy to do sentiment analysis on Twitter, especially comments in Indonesian, because the text is not structured. This study uses data from Indonesian-language tweets containing the word Mudik, the algorithm model used in this study, Naive Bayes Classifier and Support Vector Machine, is compared to get accuracy, precision, recall, and F1-score values. From this research, it was concluded that the Naive Bayes algorithm and Support Vector Machine performed well enough to predict the sentiment of tweets about Mudik on Twitter social media. Naive Bayes with an accuracy of 82% and f1-score 0.8, while Support Vector Machine with an accuracy of 87% and f1-score 0.87.