Credit card fraud has emerged as a major challenge for financial institutions due to the rapid growth of digital transactions and the increasing sophistication of fraudulent activities. Traditional fraud detection systems primarily rely on rulebased mechanisms or single machine learning models that often struggle with highly imbalanced datasets and dynamic fraud patterns. These limitations frequently lead to high false-positive rates, delayed detection, and reduced reliability of automated systems. To overcome these challenges, this study proposes a Hybrid Machine Learning Framework for Credit Card Anomaly and Fraud Detection that integrates both supervised and unsupervised learning techniques to improve detection accuracy and system adaptability. The framework combines classification algorithms with anomaly detection models to identify both known fraud patterns and previously unseen fraudulent behaviors. Data preprocessing techniques such as normalization, feature selection, and handling class imbalance are applied to improve model performance. Multiple machine learning algorithms are trained and evaluated using transaction datasets to determine optimal performance in terms of accuracy, precision, recall, and F1-score. The hybrid architecture enhances fraud detection by leveraging pattern recognition capabilities of supervised models while simultaneously detecting abnormal behavior through unsupervised learning. Experimental evaluation demonstrates that the proposed framework significantly improves fraud detection accuracy while reducing false positives compared with traditional methods. The system is designed to support real-time transaction monitoring, enabling financial institutions to detect fraudulent activities more efficiently. Overall, the proposed hybrid framework provides a scalable and intelligent solution for modern financial security systems and contributes to improving the reliability and effectiveness of automated fraud detection technologies.
AJACCM (Sat,) studied this question.