Demand Forecast Method For Build-To-Order Products Using Estimate Information As A Leading Indicator

Abstract

A method for forecasting the number of parts shipments is proposed for build-to-order products using pre-sales estimate information as a leading indicator. Since the target number of parts shipments changes irregularly and over a short lifecycle, it is difficult to forecast with conventional methods. The method uses Kalman filter to correct for the noise associated with the leading indicator and to facilitate the practical application. The method was evaluated using 78 different parts of an electronic product. Experimental result revealed that the method is useful for forecasting the target shipping number.