AI FRAUD SHIELD

Financial institutions face a critical paradox: the most accurate fraud detection models are also the most opaque, 

     making them legally undeployable under emerging global regulations. This project proposes a two-phase explainable 

     ensemble learning framework that achieves both goals simultaneously — strong fraud detection AND full regulatory 

     transparency documentation aligned to EU AI Act Articles 10 and 13.