<![CDATA[The issue of frankincense downstreaming in Indonesia is a significant concern because it has the potential to increase the economic value of local commodities and the welfare of the community, especially farmers. However, public perception of this downstreaming policy is still diverse and has not been widely analyzed scientifically, especially on social media. Therefore, this study aims to analyze the sentiment of social media users X towards the issue of frankincense downstreaming using the Naïve Bayes algorithm. The research data was obtained through a crawling process using the Twitter API with the keywords "Frankincense Downstreaming" and "Downstreaming", resulting in 1,844 tweets. The data then went through a preprocessing stage including cleaning, case folding, normalization, tokenizing, stopword removal, and stemming, leaving 1,790 tweets ready for analysis. The sentiment labeling process was carried out using a lexicon-based approach with three categories: positive, negative, and neutral. Feature representation was carried out using the TF-IDF method, then the data was classified using the Naïve Bayes algorithm. The test results show that the Naïve Bayes algorithm is able to classify sentiment well, with the highest precision in the negative class at 0.90 and the highest recall in the neutral class at 0.92. The majority of X users showed neutral sentiment towards the issue of frankincense downstreaming at 55.20%, followed by positive at 26.03% and negative at 18.77%.]]>
