A Fraud Detection Method for Online Payment Transactions Based on Deep Learning

With the rapid development of the Internet, e-commerce and internet finance have achieved rapid development, but it has also brought serious problems of online payment risk fraud. Among them, the imbalance of financial data and the accuracy of online fraud detection are the main problems. Processing the original data and optimizing the algorithm model are important means to improve the accuracy of online fraud detection of unbalanced financial data. Among them, SMOTE algorithm, deep learning model, and mixed sampling algorithm are widely used. The deep learning model improves the accuracy of online fraud detection by improving the network structure. This article proposes an online transaction fraud detection method based on deep learning and SMOTE algorithm, which synthesizes minority class samples using the SMOTE algorithm and extracts data features using deep learning models to solve the problems of data imbalance and low fraud detection accuracy. This method is expected to provide more reliable guarantees for relevant industries, assist financial institutions and online payment platforms in timely and effective fraud detection and prevention, and provide guidance and suggestions for industry behavior.