Research On Prediction Of Coupon Use Based On Xgboost Method

In recent years,China's Internet e-commerce is constantly developing and upgrading.As a major part of it,e-commerce platforms often launch various marketing campaigns.Coupons have become a major way for e-commerce platforms to launch promotions.As a common way of promotion,coupons can be used to deduct a certain amount of money when certain conditions are met at the time of settlement when the corresponding type and amount of goods are purchased within a specified time and date.By issuing coupons,we can promote and guide users to purchase corresponding products and deduct a certain amount when placing orders,so as to achieve the goal of promoting sales and increasing customer flow.Coupons are widely available both online and offline.For example,delivery coupons issued by the delivery platform,special car coupons issued by didi,and coupons issued by taobao word-of-mouth.In order to make the prediction precision and generalization to improve some,we will not only through Xgboost model forecast,and the other model to forecast,such as through GBDT Xgboost and other models and fusion,enhance the accuracy in predicting a coupon to use,integrated approach to learning,this paper combines engineering characteristics,using GBDT model with Xgboost model integration,and through the parameter tuning to improve the performance and accuracy.Experiments show that the final fusion model can improve the accuracy and generalization ability of coupon usage prediction.