A Financial Fraud Detection Model Using Artificial Intelligence

Finance, government, corporate, and consumer fraud all have broad effects on society, and as the use of new technologies increases, this problem is worsening. In the era of big data, traditional methods for detecting anomalies are slow, costly, inaccurate, inefficient, and impractical. There is a growing trend among financial institutions to adopt statistical and computational techniques in their operations. However, a comprehensive framework that integrates the entire AI lifecycle for adaptive fraud detection is still needed. This study presents, implements, and validates a five-phase Financial Fraud Detection Model (FFDM) framework. A prototype was developed and evaluated using a public dataset, comparing multiple machine learning models. The results show that the framework is operational and that an SVM model integrated into the FFDM achieved the best baseline performance for this task, demonstrating the model's practical feasibility.