<![CDATA[This study aims to analyze public sentiment toward the controversy surrounding President Joko Widodo’s academic credentials by examining user comments on YouTube. A total of 20,294 comments were collected and processed through text cleaning, normalization, tokenization, stopword removal, and stemming. Sentiment labels were assigned using a lexicon-based approach, producing positive, negative, and neutral categories. The experimental results indicate that the combination of SVM, TF-IDF, and SMOTE achieved strong classification performance, with an accuracy of 86.87%. The model demonstrated better performance in identifying negative and neutral sentiments, while some positive sentiments tended to be misclassified as neutral. Overall, this study shows that sentiment analysis based on YouTube comments can serve as an effective approach for mapping public opinion on socio-political issues in an automated and large-scale manner. Feature extraction utilized Term Frequency–Inverse Document Frequency (TF-IDF), and sentiment classification was performed using a Support Vector Machine (SVM). The model achieved an accuracy of 86.87% and a macro F1-score of 0.87, indicating that the integration of TF-IDF, SMOTE, and SVM is effective for large-scale sentiment classification of YouTube comments related to socio-political issues.]]>
Copyrights © 2025
