XGBoost-Based E-Commerce Customer Loss Prediction. (31st July 2022)

• Record Type: Journal Article
• Authors: Gan, Lin
• Other Names: Sharma Kapil Academic Editor.
• Is Part Of: Computational intelligence and neuroscience. Volume 2022(2022)
• Publication Date: 2022-07-31
• Subjects: Neurosciences -- Data processing -- Periodicals; Computational intelligence -- Periodicals; Computational neuroscience -- Periodicals; 612.80285
• DOI: 10.1155/2022/1858300
• Languages: English
• ISSNs: 1687-5265

Abstract

Abstract : In recent years, with the rapid development of mobile Internet, more and more industries have begun to adopt mobile Internet technology, provide diversified wireless services, and further expand user activity scenarios. The core of reducing customer loss is to identify potential customers. In order to solve the problem of how to accurately predict the loss of customers, this paper put forward an invented method to verify and compared the model with the customer data of an e-commerce enterprise in China. According to the research results, the improved XGBoost algorithm can effectively reduce the probability of class I errors and has higher accuracy, among which the accuracy has increased by 2.8%. The prediction effect of customer groups after segmentation was better than that before segmentation, in which the probability of the occurrence of class I errors in the prediction of core value customers decreases by 10.8% and the accuracy rate increases by 7.8%. Compared with other classification algorithms, the improved XGBoost algorithm had a significant improvement in AUC value accuracy rate and other indicators. This fully shows that the XGBoost algorithm can effectively predict the loss of e-commerce customers and then provide decision-making reference for the customer service strategy of e-commerce enterprises.