Research on Customer Value of E-commerce Data Based on Machine Learning

Abstract

Nowadays, online shopping has popularized all over the world, the e-commerce platforms have to make every effort to compete with each other. In order to survive from the fierce competition, the e-commerce platforms have to make more profits with less cost which means that high-quality customers should be the most important resource for the e-commerce platforms. Therefore, the purpose of this passage is to measure the customer value by analyzing the e-commerce data with the methods of K-Means, RFM model and logistic regression. In the experiment, the RFM model shows that the customers in different clusters have different value for the e-commerce platform, the prediction model formed by logistic regression, visualized by confusion matrix and evaluated by 4 indicators shows that only 50.4% of the customers will repurchase on this e-commerce platform in the future. So, it is necessary for the e-commerce platform to carry out the personalized strategies for the customers in different clusters.