<![CDATA[Credit card fraud is one of the major challenges in modern digital payment systems. The increasing volume of online transactions raises the potential for unauthorized use of cardholder data. This research aims to develop a robust and accurate fraud detection system by integrating two machine learning algorithms, Random Forest and XGBoost, both of which are known for their high performance in data classification. The research process begins with the collection and preprocessing of credit card transaction data, followed by model training using the selected algorithms. The model’s performance is evaluated using metrics such as accuracy, precision, recall, and F1-score. To enable real-time application, the model is implemented in a web-based system using the Python Flask framework, allowing direct integration into financial transaction environments. The need for adaptive systems that can respond to emerging fraud patterns serves as a key motivation for this study. By combining two complementary algorithms within a single web application platform, the system is expected to detect fraudulent activities quickly and accurately. The expected outcomes of this research include: (1) an optimized fraud detection model based on Random Forest and XGBoost, (2) a prototype web application developed with Python Flask for system implementation, and (3) a scientific publication describing the development and results of the proposed system. The targeted outputs are a publication in a nationally accredited journal (Sinta 4) and intellectual property registration. This research is expected to provide a significant contribution to preventing credit card fraud through the effective application of machine learning technologies]]>
