Research On Order Quantity Prediction Of O2O Business Of Chain Supermarket Based On Stochastic Forest Model

Entity chain supermarket enterprises are facing great changes in the new retail era.In the new retail era,the traditional profit model of entity chain supermarket enterprises has been unable to adapt to the development of The Times.They need to take the train of Internet platform,follow the footsteps of big data,artificial intelligence and other new technologies in the development,integrate online and offline shopping methods,and reconstruct the three elements of "people,goods and market".The great changes brought about by the restructuring are:Firstly,connecting consumers,commodities and shopping places,reshaping the business structure and comprehensively locking customers’comprehensive needs for online services and offline shopping;Secondly,the offline business mode expands online O2O business.Based on big data and cloud computing technologies,the problem of offline business data island is solved.The data island of customer information and behavior is broken,and a complete and comprehensive data chain is formed.Third,managers can easily analyze customers’ consumption habits and develop more targeted personalized services for customers,so as to improve customers’ shopping experience and make more customers stay on their own initiative to create more income.In this context,online O2O business order volume prediction is a typical application scenario of physical supermarket chain enterprises transforming to new retail.Physical supermarkets have made good use of store resources for online O2O business recommendation,and consumers have experienced the good experience of online ordering and offline delivery,forming a new generation of consumer groups.These online O2O orders have actually increased the sales of physical supermarkets and brought new growth points to enterprises.Therefore,accurate prediction of online O2O business order volume is crucial for O2O operation.The more accurate the forecast of order quantity,the more reliable the decision support of online operation business.Not only that,the accurate prediction of order quantity can also provide decision-making support for physical supermarkets to prepare goods and reasonably prepare transport capacity.Good business decisions can bring good shopping experience to customers,thus forming a virtuous cycle.In this paper,the common sales forecasting methods are comprehensively described,and the application research of supermarket order quantity forecasting is carried out based on stochastic forest algorithm model.This paper focuses on how to use the stochastic forest algorithm model to predict the order quantity of Z supermarket.This paper studies the order volume data of Z supermarket in the past two years,and explores the influences of stores and types of stores,holiday seasons,promotional coupons and even customer behaviors on the order volume from various aspects.Based on the exploration of influencing factors of order quantity,two models are selected to carry out prediction research respectively.In this paper,the principle of random forest prediction model is described in detail and the model is used for prediction.Then the principle of ARIMA autoregressive prediction model is described and the prediction is made.The accuracy index RMSE is used to compare the two models and determine which prediction model has better performance in predicting short-term order quantity.The experimental results show that the prediction results of random forest model are better than ARIMA model.This conclusion can be extended to predict the future O2O business of brick-and-mortar stores and supermarkets,which can be used as an important basis to guide the online O2O business decisions of offline stores and supermarkets under the new retail background.