Host: The Japan Society for Management Information
Due to the rapid technological innovation, discontinuation of service parts production is getting earlier. It is a few years after the release of the product that manufactures have to decide the procurement of the parts to be required for coming 10 to 20 years. Time series forecasting methods such as exponential smoothing are unsuitable for such long term forecasting in the early stage. Thus, we propose a new demand forecasting method for service parts utilizing empirical distribution of the demand. The distribution is constructed from the demand of the parts whose attributes are similar to the target part. It is confirmed that the proposed method has the ability to forecast the parts on which the conventional methods do not work.