<![CDATA[Social media has become an important tool for people to express their opinions about various institutions, including universities. This research aims to analyze Universitas Kristen Immanuel (UKRIM) using the Support Vector Machine (SVM) classification method. Data collected from social media through web scraping was processed through several preprocessing stages such as data cleaning, tokenization, and stopword removal. One of the main challenges in this study is class imbalance, which was addressed using two techniques: Synthetic Minority Over-sampling Technique (SMOTE) and Generative Adversarial Network (GAN). SMOTE was used to oversample the minority class, while GAN was utilized to generate artificial data that resembles real input. Feature extraction was carried out using the Bidirectional Encoder Representation from Transformers (BERT) model, and classification was performed using SVM. The evaluation results of the experiments show that the use of GAN to create artificial data and SMOTE to balance the classes in the data improves the classification model performance, as the SVM model can effectively distinguish negative, neutral, and positive sentiments. Keyword: SVM, SMOTE, GAN, sentiment analysis, social media]]>
Copyrights © 2025
