Analysis and mining of user behavior on e-commerce platforms based on decision trees and RFM models

User transaction behavior has driven the development of e-commerce. This article reveals the key role of the development of the e-commerce industry and user behavior in driving it through in-depth research on e-commerce sales data. Firstly, by grouping and summarizing user IDs through Pandas, a comprehensive user consumption situation was obtained, including consumption frequency, amount, etc., and user consumption behavior was visualized through visual means. On the basis of constructing effective features, Matplotlib was used to depict the number of users who receive coupons every week and day, providing a basis for the establishment of subsequent coupon distribution prediction models. By splitting the payment dates, an RFM model was constructed to analyze the purchase time period, frequency, consumption amount, and the proportion of each user's time behavior, forming a detailed user profile. Finally, the decision tree model is applied to predict the probability of coupon usage.