Perbandingan Support Vector Machine dan Naïve Bayes Terkait Kepuasan Pengguna Bus Listrik Kota Medan

The city government has introduced an initiative to use electric buses as a cleaner and more sustainable alternative. The success of a public transportation system is not only determined by the availability of fleet and infrastructure, but also by the level of user satisfaction. User satisfaction is an important indicator that reflects the extent to which the service meets users' expectations and needs. Therefore, an in-depth understanding of the factors that influence user satisfaction, as well as the ability to predict and manage them, is key in improving the quality of public transportation services, including usage. In an effort to understand and improve trolleybus user satisfaction in Medan City, an adequate predictive analysis approach is required. By using methods such as Support Vector Machine (SVM) and Naïve Bayes, we can develop predictive models that can identify patterns and trends in user data, thus enabling relevant parties to take appropriate actions to improve services. In this context, the comparison between SVM and Naïve Bayes methods will provide valuable insights into the effectiveness of each method in predicting the satisfaction of electric bus users in Medan City. Based on the comparison results, the Naive Bayes algorithm shows slightly better performance compared to the Support Vector Machine in this sentiment analysis. The accuracy value generated by applying the Naive Bayes method is 58%, while applying the Support Vector Machine method is 57%. Nonetheless, both algorithms provide valuable insights into the sentiment of Medan people towards Electric Buses.