Design And Realization Of Automatic Selection Of Bank System Transaction Path

In recent years,the emergence of various new financial payment channels has increased the complexity and uncertainty of the choice of trading routes for bank receipt systems.In order to achieve timely and efficient access to new payment channels,reduce the difficulty of system operation and merchant transaction costs,and improve the user experience,this thesis proposes a solution to the automatic selection of transaction paths.The main tasks of the thesis are as follows:First of all,it explains the knowledge of the bank card acquiring business and the basic concept of the bank receipt system.It introduces the basic concepts,development direction and application scenario of the decision tree and artificial neural network.Secondly,through the requirements analysis,the overall design plan of the bank receipt system is proposed.The program consists of three parts: the receipt transaction system,the data mining module and the forecasting service module.Third,in the design and implementation of the receipt transaction system,the design concept,implementation and operation of the framework are described in detail.In the design and implementation of the data mining module,the decision tree and artificial neural network model are constructed separately.,And compared it to the test,finally determined the artificial neural network algorithm as the automatic selection of the forecasting algorithm for the transaction path.In the design and implementation of the forecasting service module,describes the hierarchical design and interaction of the forecasting request submodule and the forecasting service submodule.Finally,the functional aspects and prediction accuracy of the system were tested.The test results show that it is feasible to use the artificial neural network algorithm model to predict the transaction path,and the function and prediction accuracy of the system are to meet the design requirements.