Support Vector based Demand Forecasting for Semiconductor Manufacturing

Accurate demand forecasting is the key capability for manufacturing organization, more so, a semiconductor manufacturer. Many crucial decisions are based on demand forecasts. In this paper, we apply support vector regression (v -SVR) with our three phased approach for modeling input vector to enhance forecasting accuracy and compare its results with other baseline forecasting techniques. Since support vector machines (SVM) have greater generalization ability and guaranteed global minima for given training data, it is believed that SVR will perform well for time series analysis. Compared to other demand forecasting techniques, our results show that the SVR can significantly reduce both mean absolute percentage errors and normalized mean-squared errors of forecasts. We demonstrate the feasibility of applying SVR in semiconductor demand forecasting and prove that SVR is applicable and performs well for semiconductor demand data analysis.