Research On The Big Data Application Of Bank Based On Marketing Response

In recent years,big data has infiltrated industries and business functions.And its size and storage capacity are growing rapidly.The high data storage in the banking industry.The updating and deepening of big data applications have brought more opportunities and innovation to the development of the banking industry.Banks' applications for big data are mainly in risk management,operational optimization,financial management and marketing.The marketing ideas and patterns of Banks are becoming more and more humane.Database marketing using big data also gradually replaces traditional marketing with its high accuracy,high efficiency and low cost.Database marketing can target the bank's target population to customers with higher responsiveness.This can reduce the bank's marketing costs,improve the return on investment and enhance the customer's user experience and satisfaction.In this paper,a foreign bank as an example,whether a bank customer will buy foreign exchange products was predicted.Based on the bank's customer database,response model and customer segmentation model was built.Prior to modeling,variables are pretreated and filtered.Compressing categorical variables were used the Weight of Evidence method,and filtering continuous variables were used the clustering method and were selected with the smallest value of 1-RSquareRatio in each category.The response model is based on the Logistic regression and the Decision Tree method to predict the customer response probability.And the two methods were compared with the ROC curve,K-S statistics and LIFT graph.In this paper,the response model of Logistic regression were relatively better than Decision Tree for fit and prediction effect.According to the customer's personal information,the various activities in the bank's performance and account transactions,customer segmentation model was built.Finally,four categories of customer groups were divided,and customer portraits were described based on the customer group to know customer characteristics and behavior preferences.it is found that the first class of customers have a higher level of education,online banking,good economic level,consumption preferences partial savings,and the high level of savings.Such customers have a higher response rate to foreign exchange products,and they can be positioned as potential customers of foreign exchange products.Finally,based on the results of customer segmentation model and marketing response model,the appropriate marketing strategies were developed to the bank,and potential customer groups was dug out.The purpose of saving costs and improving the marketing response rate were achieved.