A Consumption Analysis of Sugar-sweetened Beverages Ordered Online for Home Delivery:An Application of Machine Learning
Sugar-sweetened beverages have caused worldwide health concerns due to their contri-bution to high calorie intake and chronicle diseases.Against the backdrop of the rapid development of online food delivery platforms,this paper utilizes sales data from delivery services in six Chinese cities and employs machine learning methods to identify consumption patterns of sugar-sweetened beverages.Additionally,the paper analyzes sales variations in simulated scenarios such as price discounts and sugar taxes.The findings of the paper are threefold:First,neural networks within supervised machine learning models demonstrate greater applicability for predicting consumption trends compared to tradi-tional econometric regression models.Second,model predictions indicate that a 10％price reduction leads to a 0.49％increase in total sales,suggesting significant price elasticity of demand for sugar-sweetened beverages.Conversely,a 10％price increase results in a 0.17％decrease in total sales vol-ume,implying that the utilization of taxation as a strategy to reduce the consumption of these beverages is viable.Third,there are spatial and temporal variations in the sales of sugar-sweetened beverages on delivery platforms,with higher sales volumes in the Yangtze River Delta region compared to other areas,and weekday sales surpassing those of weekends,with afternoons being the peak period.This research is significant for establishing pathways and mechanisms to reduce excessive consumption of sugar-sweet-ened beverages,guiding consumers towards healthier dietary choices,and promoting sustainable devel-opment in the food delivery service industry.
NETL
NSTL
万方数据
