Analisis Sentimen pada Aplikasi Kredivo dengan Metode Algoritma C4.5 Berbasis Forward Selection

Abstract

The increasing use of digital financial applications such as Kredivo highlights the need to understand user perceptions and satisfaction. This research aims to analyze user sentiment towards the Kredivo application available on the Google Play Store using text mining techniques and the C4.5 classification algorithm enhanced by forward selection for feature optimization. User review data were collected and processed through a series of steps, including text preprocessing, feature extraction, and sentiment classification into positive and negative categories. The evaluation of the model was conducted using performance metrics such as accuracy, precision, and recall. The results show that user sentiment is predominantly negative, primarily due to technical issues with the application and poor customer service. Some users also expressed concerns regarding app features and data security. The sentiment classification model achieved an accuracy of 76%, precision of 50%, and recall of 14%. These findings suggest that while the model performs well in general classification, improvements are needed in detecting positive sentiment. This study contributes to the understanding of user experiences with financial technology services and can support application developers in enhancing user satisfaction and service quality.