Research on E-commerce Customer Churn Prediction Based on Improved Value Model and XG-Boost Algorithm

摘要

In recent years, with the development of Internet technology, the market competition is fiercer, the cost of acquiring new customers is increasing, and the cost of maintaining old customers is far less than the cost of acquiring new customers. Most companies are trying to market precisely through customer segmentation in order to reduce the rate of customer churn. Aiming at the customer characteristics of social network e-commerce, this paper builds a customer value model that integrates the value of social network to help companies subdivides the customer accurately. Then we use the machine learning algorithm XG-Boost to predict the churn of customers before and after the subdivision. The research found that the prediction accuracy is higher after customer segmentation. In addition, the XG-Boost algorithm is more advantageous than other algorithms.