| Commodity exchange is a basic social activities,all over our lives,e-commerce makes the scale of this social activity surge,followed by demand forecast for the significance of the electric business more and more important.At present,the research direction of forecasting system at home and abroad is to improve the dynamic prediction accuracy of the model and the combination forecasting of the model.Combining the forecasting model with inventory planning is the development direction of warehouse planning.This paper is to predict the model and inventory planning combined to predict and sub-warehouse planning method.First of all,based on the correlation analysis rule set by the author,the transaction data of Alibaba electronic business platform are excavated to obtain the effective features,and the model is used to predict the demand of these goods in the next two weeks.By comparing the forecasting performance of gray forecasting model,neural network model,multiple regression model and stochastic forest model,the random forest model is selected as the regression model to forecast the demand of commodity in the next two weeks,and a relatively accurate prediction result is obtained.Then,based on the prediction results,the optimal storehouse planning of the newsboy model is established on the basis of residual analysis.On the basis of the solution of the model,combined with the observation of commodity trading in life and the experience of purchasing certain commodities,adopting the method of empowerment to optimize the result and get the result closer to the real demand.Finally,the paper proposes the sub-warehouse planning based on the principle of minimum cost.In this paper,the characteristics of a large number of features of the electronic business transaction data extraction model established by the help of electric business enterprises to effectively predict the demand and thus develop a lower cost sub-warehouse planning. |
PDF Full Text Request