A Demand Forecasting Method Based on Likelihood Estimation of Characteristic Coefficients and Retail Business Rules

Abstract

The invention discloses a demand prediction method based on characteristic coefficient likelihood estimation and retail business rules. Firstly, the invention first obtains store traffic and external factor data such as weather, temperature, holidays, etc., and uses a linear regression method to obtain an estimated amount of future turnover; Secondly, the characteristics of the week and the corresponding sales ratio of the week are constructed, and the stores are clustered after defining the similarity of the stores based on the Euclidean metric rule. By merging the weeks with similar sales ratios, the week group features are constructed; then, the weighted temperature and The average value of recent sales is added to the feature set, and the daily sales distribution is obtained by using the GLM model; finally, the daily delivery demand is calculated based on the store's display requirements, order restrictions and other rules to complete the retail demand forecast. The invention constructs model features based on business considerations, establishes a commodity demand model in the retail industry, and greatly improves the operating efficiency of retail stores.

Description

technical field [0001] The invention relates to the field of supply chain automation management engineering, specifically, a demand prediction method based on characteristic coefficient likelihood estimation and retail business rules serves terminal operation and management. Background technique [0002] In today's supply chain automation management engineering field, whether it is actual business practice or theoretical research, optimizing supply chain inventory to meet customer needs has always been one of the core businesses of enterprises, reducing inventory investment while ensuring retail shelves and distribution The center stocks just the right amount of merchandise. [0003] Most of the existing intelligent management platforms are general-purpose platforms, and the general-purpose solutions are not customized for the details of the enterprise's business. Enterprises are still required to control the terminal needs and deployment. The amount of suggestions provided ...

AI Technical Summary

Problems solved by technology

[0003] Most of the existing intelligent management platforms are general-purpose platforms, and the general-purpose solutions are not customized for the details of the enterprise's business. Enterprises are still required to control the terminal needs and deployment. The number of suggestions provided by the platform system also has large deviations in estimation, and the calculation plan is not comprehensive. Failure to deal with more business restrictions and business changes in terminal retail

Images