Based On Neural Network Bank Card Anti-fraud System Model And Its Empirical

With the bank card is getting popular and its market is getting mature, how to predict and identify new types of fraud becomes particularly important. Abailable products abroad and most of the theories are based on the special environment. Many small and medium-sized banks do not have enough conditions to run these products. This paper presents a new practicial model to meet the modern financial theory. It's expected that the research result will be helpful in developing the bank card anti-fraud model.This paper takes an in-depth study of the existing anti-fraud system models, and analyzes the obstacles of building an anti-fraud system in the small and medium-sized banks. By importing neural network technology, a neural network-based bank card anti-fraud system model is provided, which is applicable for the status of small and medium-sized banks. This paper describes the detail structure of the anti-fraud system model, the operating mechanism and the algorithm of its core subsystems. By appling the model to a bank card's applications, the algorithm and the model in the core subsystem's is proven to be reasonable.This paper investigates the existing bank card anti-fraud system's development process. And then compares two anti-fraud models prevailing in China and international: based on artificial intelligence model and rule-based model, further analyzes the advantages and disadvantages between the models. The obstacles of building an anti-fraud system in the domestic small and medium-sized banks are identified. A neural network-based bank card anti-fraud system model for small and medium-sized banks is presented. The bank card anti-fraud system model is composed of three subsystems: the core detection subsystem, the data acquisition subsystem, the management subsystem. In the core detection subsystem, its key technology and algorithms is studied and is applied in the model anti-fraud system in a bank debit card online payment system. Some experiments are used to test its operation mechanism and algorithms on the core detection subsystem. The results demonstrate that using neural network model to simulate user's consumption can detect fraudulent transactions. Finally the paper summarizes the research and gives a view of future direction of the anti-fraud model.The main contribution of the paper is as follows:1. Compares the prevailing anti-fraud models and the status of small and medium-sized banks in China, proposes a neural network-based bank card anti-fraud system model which is applicable for the small and medium-sized banks.2. Studies the neural network model which is useful for behavioral patten's detection in the anti-fraud model. A practical application of the banking online system proves that the neural network model's design is reasonable.