Penerapan Model Word Embedding IndoBERTweet pada Metode Support Vector Machine untuk Klasifikasi Opini Publik di Media Sosial X

Abstract

This study aims to classify public sentiment regarding the redenomination of the Indonesian rupiah on social media platform X using a combination of Support Vector Machine (SVM) and IndoBERTweet word embedding. The main challenge in social media sentiment analysis lies in unstructured text, informal language usage, and contextual ambiguity. Therefore, an approach capable of capturing contextual meaning while maintaining high classification accuracy is required. This research employs a quantitative approach, including data collection through crawling, text preprocessing, data labeling, feature extraction using IndoBERTweet, and classification using SVM with a Radial Basis Function (RBF) kernel. A total of 1,014 tweets were collected and refined into 370 labeled data consisting of positive and negative sentiments. The results show that the proposed model achieves an accuracy of 82\%, precision of 83\%, recall of 82\%, and F1-score of 82\%. These findings indicate that the integration of IndoBERTweet and SVM effectively captures contextual semantics in Indonesian social media text and improves sentiment classification performance. Furthermore, the analysis reveals that the majority of public opinions tend to be positive toward the redenomination issue. This study is expected to contribute to the development of machine learning-based sentiment analysis and support more responsive policy-making based on public opinion