ANALISIS SENTIMEN PENGGUNA SOSIAL MEDIA X TERHADAP PEMILIHAN ANGGOTA KABINET MERAH PUTIH DENGAN METODE SVM (SUPPORT VECTOR MACHINE)

Abstract

Social media has become a primary platform for people to voice their opinions on political issues, including in response to the formation of the Red and White Cabinet. The diverse public expression spread across platforms such as Twitter (now X) provides a potential data source for computational analysis. This study aimed to analyze public sentiment towards the cabinet formation through a machine learning approach using the Support Vector Machine (SVM) method. The research process included text preprocessing stages such as case folding, cleansing, tokenizing, stopword removal, and stemming, followed by feature extraction using TF-IDF. Model evaluation results show that SVM was able to classify negative, neutral, and positive sentiment with an accuracy of 96% on the training data and 88% on the testing data. The best performance was shown in the classification of negative sentiment (f1-score 0.92) and positive (f1-score 0.88), while neutral sentiment showed low performance (f1-score 0.47) due to data imbalance. Sentiment analysis of public responses revealed that 56% of posts were negative, 41.5% were positive, and only 2.5% were neutral. The dominance of negative sentiment reflects public criticism or dissatisfaction with the cabinet lineup. This research demonstrated the potential of SVM as a reliable method for understanding public opinion on social media, while also highlighting the importance of handling imbalanced data and developing models for more accurate classification in the future.