Analisis Sentiment Terhadap Barak Militer pada Media Sosial X Menggunakan Metode Support Vector Machine (SVM)

Abstract

The Government of West Java Province initiated a military barracks program aimed at building character and restoring the identity of troubled adolescents through a semi-military approach. This study aims to analyze public sentiment toward the program on social media X by applying the Support Vector Machine (SVM) method, optimized using the Synthetic Minority Over-sampling Technique (SMOTE) to address data imbalance. The data were collected through a crawling process, resulting in 7,039 tweets after duplicate filtering and removal of neutral data. The preprocessing stages included cleaning, case folding, tokenizing, normalization, stopword removal, and stemming. Sentiment labeling was performed using a lexicon-based approach, resulting in 6,596 positive tweets and 443 negative tweets. The SVM model was trained using 80% of the data and tested using the remaining 20%. The evaluation results show that the model achieved an accuracy of 96%, with 45% negative precision, 98% positive precision, 48% negative recall, 98% positive recall, 46% negative F1-score, and 98% positive F1-score. Visualization using a word cloud further supports the finding that public opinion in the dataset tends to be positive, particularly regarding perceptions of improved discipline and adolescent character development. The contribution of this study lies in developing sentiment analysis of the military barracks program through the application of SVM combined with SMOTE and contextual interpretation of public opinion patterns. The findings provide not only an overview of classification performance but also identify public support for discipline and character development alongside criticism of potential stigmatization and coercive approaches to adolescent development.