XGBoost Based Strategic Consumers Classification Model on E-commerce Platform

Abstract

The strategic consumption behavior that manifests as "holding money and delaying the purchase" is an important factor affecting the profits of e-commerce platforms. Studies have shown that ignoring the strategic consumption behavior of users will bring up the losses that are equal to 30% profits. After selecting strategic consumers, companies can make targeted pricing and marketing for this particular group of people, thereby reducing the waiting time for strategic consumers. Therefore, this paper proposes a strategic consumers classification model for e-commerce platforms based on the XGBoost model. After through effective processing of JD Mall user behavior data, we build characteristics consistent with strategic consumer behaviors and use XGBoost to train and tune. The accuracy of the classification model reached 89.59%. The effect of XGBoost has achieved a better classification result than that of other classification models, thus this model can provide references for personalized recommendations and precise marketing in practical applications and increase corporate's profits.