Analisis Sentimen Terhadap Ulasan Aplikasi MyPertamina pada Google PlayStore dengan Menggunakan Algoritma K-Nearest Neighbor dan Support Vector Machine

Abstract

In the era of digital transformation, technology-based applications are key to improving the efficiency and transparency of public services. MyPertamina, an application from PT Pertamina (Persero), is designed to support fuel subsidies and facilitate digital transactions. However, the app has faced obstacles, such as complaints related to performance, user interface, and technical issues. This research aims to analyze the sentiment of MyPertamina user reviews on Google Playstore using K-Nearest Neighbor and Support Vector Machine algorithms. Review data is collected through web scraping techniques and classified into positive and negative sentiments. The research methodology used is Knowledge Discovery in Database. Of the 24,777 review data about the MyPertamina application, it resulted in 21,391 clean data after preprocessing, then the data labeling process using IndoBert. Based on the labeling, there are 8,849 positive class reviews, and 12,542 negative class reviews. This research shows that the MyPertamina application gets more negative classes than positive classes. The results of applying the K-Nearest Neighbor algorithm resulted in an accuracy of 85% and the results of applying the Support Vector Machine algorithm resulted in an accuracy of 91%.