Demand forecasting for footwear products using wavelet transform and Artificial Bee Colony algorithm optimized Polynomial Fitting

Abstract

Product life cycle management plays a crucial role in the footwear products demand forecasting. However, it is difficult to predict the recession point of the product life cycle curve. This paper proposes an integrated forecasting system where wavelet transforms and Polynomial Fitting based on Artificial Bee Colony algorithm are combined for footwear products demand forecasting. First, the method of wavelet transform using the one dimensional discrete wavelet is applied to decompose the sales data and thus eliminate the noise. Second, the Polynomial Fitting is employed to simulate the product life cycle function of footwear products. Third, the Artificial Bee Colony algorithm is utilized to optimize the parameters of Polynomial Fitting. Finally, real-world evaluation on a Chinese shoes and apparels retailer demonstrates that the proposed system is highly promising. At the same time, the simulation results also show that the demand distribution during the same period are basically the same between similar stores, which illustrates good practice of this demand forecasting method.