<![CDATA[Social media, particularly YouTube, has become a crucial platform in shaping public opinion regarding national economic issues. This study aims to analyze public sentiment towards Indonesia's economic policies through YouTube comments using a machine learning approach. The dataset consists of 1,637 comments, divided into 1,309 training data and 328 testing data, with four sentiment categories: Positive, Negative, Neutral, and Mixed Sentiment. The Support Vector Machine (SVM) LinearSVC algorithm was implemented alongside TF-IDF feature extraction. The results indicate that the SVM model achieved an accuracy of 85.37% with an efficient training time of 0.04 seconds. The sentiment distribution was dominated by Neutral (75.32%), followed by Positive (14.42%), Negative (7.15%), and Mixed Sentiment (3.12%). The best performance was achieved in the Neutral category with a precision of 0.87 and a recall of 0.99 (F1-score 0.93). However, the model demonstrated significant weaknesses due to severe class imbalance: the model completely failed to classify Mixed Sentiment (F1-score 0.00) and showed low performance on Negative (Recall 0.25). The majority of the public exhibited a wait−and−see attitude toward economic policies, indicating the maturity of economic literacy among the Indonesian society in responding to national issues.]]>
Copyrights © 2025
