基于改进机器学习算法的电商用户购买预测
首发时间:2019-03-22
摘要:近年来,电商平台陡增的巨额订单量容易导致快递爆仓,也极大地超过了物流系统的运载能力。本文采用现有的数据挖掘技术分析用户的购买习惯,预测用户的购买情况,对于预测购买的用户采取提前发货的策略,这样达到合理降低运输及配送高峰的目的。并在经典机器学习算法的基础上,建立基于bagging集成学习的混合模型,采用电商数据进行仿真,仿真结果表明改进后的算法具有更好的预测能力,能够满足基本预测需求。
For information in English, please click here
Buying Prediction of E-commerce Users Based on Improved Machine Learning Algorithms
Abstract: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.
Keywords: machine learning hybrid model prediction Integrated Learning
基金:
引用
No.****
同行评议
勘误表
基于改进机器学习算法的电商用户购买预测
评论
全部评论0/1000