Public Sentiment Analysis of Telkom University: A Comparative Study of SVM and Decision Tree Models

In the digital era, online platforms, especially social media, have become the main channel for voicing opinions and perceptions on various subjects, including educational institutions such as Telkom University (Tel-U). This study aims to analyze public sentiment towards Tel-U using advanced machine learning techniques, namely Support Vector Machine (SVM) and Decision Tree algorithms. The dataset consists of public opinions meticulously categorized into positive, negative, and neutral sentiments. The results of the analysis show that positive sentiments are mainly related to the quality of education, outstanding faculty, and technological innovation, while negative sentiments focus on traffic jams, administrative issues, and education costs. Neutral sentiments, on the other hand, contain more descriptive information without strong emotional undertones. From the evaluation results, the SVM model showed superior performance compared to the Decision Tree, with an impressive average AUC-ROC score of 0.96 compared to 0.80 for the Decision Tree, and an accuracy of 84% compared to 72%. This clearly indicates that the SVM model is significantly more effective in accurately classifying public sentiment towards Tel-U. Moreover, these findings strongly suggest that machine learning models can serve as powerful and reliable tools for educational institutions to monitor and respond to public opinion dynamically and efficiently. This study provides valuable insights for Tel-U to better understand public perception and to identify specific areas that need improvement.