Genetic Optimization of BP Neural Network in the Application of Suspicious Financial Transactions Pattern Recognition
Jun Tang, Lei He · 2012
We explore and analyze the chaotic properties of the financial data, conduct classification learning in the financial transaction data. In this way, we are able to excavate the pattern and rule of customer transaction behavior, and isolate suspicious financial transactions. Using MATLAB to implement the programming of BP, we propose a genetic optimization of BP neural network to improve the defect of BP, which includes slow convergence and falling into local optimum easily. By using genetic algorithm to optimize the BP network, we are able to select the weight coefficient in a better way. Our experiments show that the optimized BP neural network function has a better predictive output.