| Summary: | 碩士 === 元智大學 === 管理碩士在職專班 === 107 === The global market continues to become fragile and volatile. Enterprises must rely on the overall supply chain members to communicate and cooperate with each other to enhance competitive advantage. Supply processes continue to mature and develop faster than the demand. Therefore, there is a large gap to get in the redefinition of the demand forecasting processes to become a demand-driven strategy than any other area of the supply chain. The current development of supply chain manage methods often ignores the variation of information flow and brings about the problem of "Bullwhip Effect". Demand forecasting, which combines information on demand, procurement, inventory, sales, etc., has a great impact on the production. Collaborative Planning Forecasting and Replenishment (CPFR) consider to the most effective way. How to address demand forecasting this area effectively will be the highest priority in the next few years, and artificial intelligence will be the core key to implementation.
This study uses externally available information for the demand forecast of manufacturing companies, taking into prosperity monitoring data and industrial production data. The Group Method of Data Handling is use to construct different demand-oriented forecasting models, and evaluate accuracy by MAD, MAPE and RMSE. This study successfully used the relevant attributes to predict the sales of the next phase. This amount can also use to measure the relevant possible sales volume, establish a demand-forecasting model with external data and good corporate identification, and enable all employees to quick use, economize labor and time.
|
|---|
