Sales forecast for pickup truck parts: A case study on brake rubber

Abstract

In this paper we address sales forecasting of brake rubber for 1600 pickup truck manufactured by Iran Khodro Co. To this end, we use two different methods named Neural Network (NN) and regression model. Further, we develop two types of neural networks, one general network and a set of monthly networks. Results reveal that when data are nonlinear and chaotic, traditional models like regression are less likely to be useful. In these cases we can use nonlinear models like neural networks. It is shown that general network is not a useful tool for forecasting sales of brake rubber, whereas monthly networks are accurate and useful for this purpose.