Design And Implementation Of Bank Marketing System Based On Improved Integrated Neural Network Algorithm

With the advent of the era of big data,machine learning technology continues to develop.For commercial banks,it is difficult to use the traditional product-centric marketing methods to accurately locate customer preferences,and it is increasingly difficult to improve the competitiveness of banks in the market.The new bank marketing strategy emphasizes the need to form long-term stable relationship with existing customers,make full use of big data resources to accurately analyze customer needs,classify customers before product marketing,predict whether customers will order financial products or not,and market different financial products to different customers.Therefore,this thesis studies the BP neural network algorithm in machine learning,optimizes the BP neural network algorithm,and designs and implements a bank marketing system for targeted marketing.The main work of this thesis is as follows:1.There is a high degree of uncertainty about the network architecture selection method of traditional BP neural network.This thesis uses the gravitation search algorithm to optimize the network architecture.To solve the problem that BP neural network is easy to overfit,the Bagging integrated learning algorithm is used to integrate the BP neural network optimized by gravitation search algorithm.The Bank Marketing dataset is used to simulate the algorithm,and it is compared with the network architecture selection method based on empirical formula,genetic algorithm and simulate anneal.The experimental results show that the method of optimizing BP neural network architecture based on gravitation search algorithm reduces the MSE of BP neural network,and the number of iterations required for the algorithm to converge to the optimal network architecture is less.A BP neural network integrated with Bagging algorithm and Adaboost algorithm is used for comparative experiments.The experimental results show that the BP neural network integrated by Bagging algorithm can significantly reduce the over-fitting risk of the algorithm and enhance the generalization capability of the algorithm.The BP neural network integrated by Adaboost algorithm does not improve the generalization performance of the algorithm.2.This thesis designs a banking marketing system for marketing a fixed-term-deposit product,and uses the improved integrated BP neural network algorithm as a predictive model.The system is mainly divided into three functional modules,which are user management,project management and marketing management.Among them,the marketing management function is the core part of the system.This module can select the marketing tasks,use the forecasting model to classify the customers,select the customers who the system thinks will order the products,and follow up marketing according to customers' previous marketing results.The architecture design,the database design,the main process and the design of each functional module are given.3.The Spring Boot framework is used to implement the various functional modules in the marketing system,and the Bank Marketing dataset is used to test the bank marketing system,and the test results meet the system design requirements.Compared with the success rate of traditional marketing,the results show that the marketing method based on BP neural network reduces the marketing risks of traditional marketing methods.