基于改进机器学习算法的电商用户购买预测

In recent years, the huge order volume of e-commerce platform increases sharply, which easily leads to express warehouse explosion and greatly exceeds the carrying capacity of logistics system. In this paper, the existing data mining technology is used to analyze the purchase habits of users, predict the purchase situation of users, and adopt the strategy of early delivery for the users who predict purchase, so as to achieve the purpose of reducing the peak of transportation and distribution reasonably. On the basis of classical machine learning algorithm, a hybrid model based on bagging ensemble learning is established. The simulation results show that the improved algorithm has better prediction ability and can meet the basic prediction needs.