<![CDATA[Access by KAI is the primary digital platform used by the Indonesian public for train ticket reservations. Although thenumber of application downloads continues to increase, the application still receives numerous negative reviews onGoogle Play Store, primarily related to system performance issues. The large volume of user reviews makes manualanalysis inefficient, highlighting the need for an automated approach to identify the main problems experienced byusers. This study aims to analyze the dominant complaint topics in Access by KAI user reviews using a Text Miningapproach with the Latent Dirichlet Allocation (LDA) method. The dataset consists of 2,000 recent low-rated userreviews (1- to 3-star ratings) collected through web scraping techniques. The research stages include data collection,text preprocessing (cleaning, case folding, stopword removal, and stemming), corpus construction, and topic modelingusing LDA. The results identify four dominant complaint topics: (1) delayed payment verification, (2) applicationinstability (crashes/errors) on Android devices, (3) malfunction of application features after software updates, and (4)difficulties accessing ticket purchasing services during peak demand periods (ticket wars). The identified topics providevaluable insights for application developers to prioritize improvements in system stability and the reliability of paymenttransaction services.]]>
Copyrights © 2025
