Study And Implement On Dynamic Sales Forecasting Technologies For The Garment Industry In M-Commerce Enviorenment

The distribution network of China's garment industry always has three characteristics: multi-modes, multi-areas and multi-levels, which lead to varies information from different fields existing in the network and thus it's more difficult to make a distribution plan. Meanwhile, because of clothing product's short lifecycle, seasonal, regional and epidemic characteristics, the clothing enterprises must build the ability of rapid reaction to market changes. In such a situation, how to make a rapid and accurate sales forecasting becomes to be the key of distribution decision in the supply chain management of the modern garment industry. The article's object is to improve the accuracy and rapidity of sales forecasting of garment industry. By analysing the present status & development trend of technology for clothing sales forecasting, a method based on the improved adjusted model is forwarded. Combining with relative technology of M-Commerce, a dynamic sales forecasting model is built, related system is developed and its application and verification is implemented in the real garment enterprise.This article mainly focuses on the study in several aspects as below:1. Sales forecasting method for the garment industry based on the improved adjustment model: Firstly, an optimized ART model is built to improve traditional AR model, the feasibility of error prediction modeling is also demonstrated. Then main influential factors of the sales of clothes are introduced to build a multivariable error prediction model using neural network. This method, which provides a dynamic close-loop forecasting method with feedback error, can improve the application of the traditional adjustment model to clothing sales forecasting.2. Sales forecasting method for the garment industry based on mobile terminals: to build a dynamic forecasting model is put proposed. This technology collects & exchange & display the real time distribution data rapidly using mobile terminals. Aiming at the charicterictics of short-term clothing sales forcasting, real-time data is introduced to improve the improved adjustment forecasting model. So a multi-stage dynamic forecasting model is built, it comprehensively improve the accuracy and rapidity of clothing sales forecasting.3. Implementation of the dynamic and accuracy sales forecasting system for the garment industry and Verification in the real garment enterprise: The objective design and implementation scheme of the system are proposed. Some technical difficulties are conquered, such as WebService, mobile page developing and forecasting model building method etc. The application and verification is implemented in the real garment enterprise. The result shows that the system greatly improved the quality of sales forecasting and supplied a better decision support for supply chain management on garment industry.