Aftermarket demands forecasting with a Regression-Bayesian-BPNN model

Abstract

The rapid development of automobile industry in China promotes the stable growth of the automotive aftermarket. For optimizing supply chain operations and reducing costs, it is critical for a company to forecast the demands for auto spare parts in the future. This paper proposes an improved Regression-Bayesian-BBNN (RBBPNN) based model to realize the demands forecasting. Compared with a classic ARMA model, the proposed RBBPNN model has higher accuracy and better robustness. These advantages are illustrated through the case study with the real sales data of a 4s shop in Shanghai.