Innovative Fraud Detection Model
This case study introduces a new innovative fraud detection model. When the rule-based and machine learning anti-financial fraud model is learning complex serialized transaction features, the effect is below expectations. At the same time, a standalone deep learning method is also showing limitations in the learning of features within a single transaction. The National Engineering Laboratory for E-commerce and E-payment, ZhongAn Technology, and Intel proposed the innovative GBDT→GRU→RF sandwich-structured fraud detection model framework, which is able to overcome the inadequate learning of serialized transaction features and single transaction features via a multilayered learning approach. Intel and partners made innovative use of a multilayered deep learning method to boost the performance of the anti-financial fraud model. The system has been tested and the results show that the solutions are feasible. It also demonstrates the further applications and innovation of advanced technologies, such as deep learning, in the financial industry.