<![CDATA[Abstract Marapthon is a live streaming phenomenon with a Subathon format that runs for several days, where the duration of the stream is determined by the amount of audience donations. This phenomenon has been growing and attracts high levels of interaction between streamers and viewers, leading to various X user opinions, both positive and negative. This study aims to analyze user sentiment toward Marapthon live streaming on the social media platform X using the Naïve Bayes method. The data were collected through a Web Scraping technique, resulting in 1,695 tweets. The collected data then underwent several stages, including preprocessing, data labeling, data splitting, feature weighting using TF-IDF, and classification using the Naïve Bayes algorithm. The results show that there are 673 positive sentiment data and 1,016 negative sentiment data. The model evaluation achieved an accuracy of 71.89%, indicating that the model has fairly good performance in classifying X user sentiment. Overall, the findings indicate that user sentiment toward the third season of Marapthon live streaming tends to be negative, although a considerable number of users still express positive opinions.]]>
Copyrights © 2026
