Analysis Of Bank Customer Portraits Based On Data Mining

With the the Internet and big data are getting hotter,all kinds of huge data have come into our lives.Every customer is a very important presence for the company,which bring huge transaction record data to its customers.For banks,customers' deposit and loan records,the number of products used,and the activity with banks are all very important and meaningful.The most important task is how to make full use of these big data to analyze the current operating status of the bank,and in the actual production and operation,use the limited resources to maximize their return.Based on big data,customer portraits can be used to achieve a comprehensive understanding their risk preference,consumption,and spending power for customers by tagging their transaction behaviors,and then provide different services and marketing for different customers.In this thesis,the transaction data of a European bank in actual operation is used as the main research object,and the relevant variables generated by it are used as indicators to establish a customer segmentation portrait and a customer churn portrait.Firstly,the data are cleaned through data preprocessing,and the customer segmentation portrait model is established by using the K-Means clustering method.Based on the decision tree,KNN algorithm and SVM algorithm,a combined churn prediction model is used to establish the customer churn portrait model.The analysis results show that from the perspective of customer segmentation,customers are divided into high-value customers,middle-value customers and low-value customers.According to the characteristics of each type of customer,corresponding service strategies and marketing strategies are proposed.From the perspective of customer churn,the customers are divided into loyal customers,churn customers and lost customers,and corresponding retention strategies are proposed according to the characteristics of each type of churn.For the customers who are unlikely to retain or whose retention value is not high,it is better to give up them.After completing the two portraits,these customers were further divided into nine categories including high-value loyal customers,medium-value vulnerable customers and low-value lost customers,and these nine categories were divided into three levels.