基于机器学习的金融客户购买行为预测  

Machine Learning-Based Prediction of Financial Customer Purchase Behavior

作　　者：程宇轩 Cheng Yuxuan(Fuzhou University,Fuzhou 350002,China)

机构地区：[1]福州大学,福建福州350002

出　　处：《黑河学院学报》2023年第10期52-56,共5页Journal of Heihe University

摘　　要：金融客户的增长为金融平台带来经济效益,但也带来了新问题:如何有效投放商品信息和提供精准服务来满足个性化和定制化的消费需求。面对海量数据,利用数据挖掘技术和机器学习算法分析和预测客户的购买需求成为关注的课题。利用招行客户的个人信息、信用卡消费和掌上生活APP上的操作日志数据研究客户的优惠券购买行为。根据购买行为的形成机制和受影响因素,运用机器学习分类算法建立购买预测模型,掌握客户对掌上生活APP上优惠券的购买意向。使用XGBoost和LightGBM算法进行样本训练,通过AUC评分衡量模型的泛化能力,通过准确的商品筛选和快速的推送服务平台为客户提供个性化的商品推荐,既节省时间,又提升金融机构竞争力和客户购物体验。The growth of fi nancial customers has brought economic benefi ts to fi nancial platforms,but it has also brought new problems:how to eff ectively release product information and provide precise services to meet personalized and customized consumption needs.In the face of massive data,it has become a topic of concern to analyze and predict customers’purchase demands by using data mining technology and machine learning algorithms.This paper studies the coupon purchasing behaviors for customers of China Merchants Bank in terms of personal information,credit card purchases and operation log data on the Palm life APP.The research uses XGBoost and LightGBM algorithms for sample training,and measures the generalization ability of the model through AUC scoring.Through accurate product screening and fast-push services,the platform can provide personalized product recommendations to customers,which will save time,improve the competitiveness of financial institutions,and enhance customers’shopping experience.

关 键 词：购买预测 LGBM XGBoost 二分类 决策树 

分 类 号：F830[经济管理—金融学]