ANALISIS SENTIMEN ULASAN APLIKASI ”ACCESS BY KAI” MENGGUNAKAN ALGORITMA MACHINE LEARNING

The trains have become one of the most popular modes of transportation in Indonesia. In line with technological advancements, PT KAI launched an application called "Access by KAI," which has received various ratings on the Google Play Store. This research aims to analyze user sentiment towards the "Access by KAI" application. The study employs the SEMMA methodology (Sample, Explore, Modify, Model, and Assess) as a framework, utilizing a combination of machine learning algorithms and lexicon-based approaches. Data collection was conducted through web scraping user reviews from the Google Play Store from January to June 2024, resulting in 9,124 reviews. The best model for sentiment classification in this study is Logistic Regression, achieving an accuracy of 84%, followed by the Random Forest algorithm with an accuracy of 78%, and Naive Bayes with an accuracy of 73%. The results indicate that negative sentiment predominates, suggesting that the "Access by KAI" application requires improvements, particularly in areas that have generated user complaints. Words such as "difficult," "login," "wrong," "payment," and "failed" reflect user frustration related to the login and payment processes, leading to user dissatisfaction. This research is expected to provide insights into user opinions regarding the "Access by KAI" application.