Fast moving product demand forecasting model with multi linear regression

Abstract

Accuracy of demand forecasting greatly influences the performance of the supply chain system which ultimately has a direct impact on the business perfomance. Accurate forecasting will be able to utilize company resources efficiently. However, in practice many companies admit that their forecasting process is not going as well as they expected. Most companies only use historical data to forecast future demand. Whereas past demand data is not enough to be used as the basis for future forecasts. Therefore it is necessary to build a model that is able to accommodate this phenomenon. This study proposed a multiple linear regression forecasting model for fast moving product. The independent variables used are climate, promotion, cannibalization, holidays, product prices, number of stores, population and income that always change over time. The results show that the proposed multiple linear forecasting model is more than three time more accurate than company forecast.