<![CDATA[The establishment of Universitas Republik Indonesia (URI) by the Indonesian government in July 2026 triggered diverse public reactions on social media, particularly YouTube, given its organizational structure has no precedent in the national higher education system . This study analyzes public sentiment toward the policy based on YouTube comments and evaluates the performance of the Naive Bayes algorithm in classifying that sentiment . Comments from fifteen relevant YouTube videos were collected, labeled using the InSet Lexicon, represented using TF-IDF, and classified using the Complement Naive Bayes algorithm . After preprocessing, 1,552 clean comments were obtained, comprising 58.2 percent negative, 27.6 percent positive, and 14.2 percent neutral sentiment . The classification model achieved an overall accuracy of 64.3 percent, performing best in recognizing negative sentiment, while most misclassifications occurred between the positive and negative classes due to lexical similarity . The dominance of negative sentiment indicates public concern over budget transparency and institutional overlap, while positive sentiment reflects hopes for expanded free education access . This study provides an initial map of public perception toward a newly launched education policy and confirms the relevance of lexicon-based and Naive Bayes approaches for rapid public opinion mapping by policymakers]]>
Copyrights © 2026
