Manufacturing Cost Prediction in the Presence of Categorical and Numeric Design Attributes

Abstract

Manufacturing processes require not only physical operation capabilities but also non-physical management policies. When designing a new product or manufacturing a customer’s new unique design, the focal point is to establish a price which maximizes customer value while still being profitable. Since an irreversible and large amount of capital is tied up in production elements, estimating manufacturing costs accurately is critical. Therefore, final decisions about the product price should be based on analytical approaches, instead of intuitive expectations. Poorly established product prices that are a function of product cost may cause two unfavorable consequences: (1) A potential loss of profit due to the gap between the expected cost and the actual cost, (2) A loss of customers and goodwill due to higher prices than necessary. In this research, we investigate ways of using clustering and spline methods to predict the manufacturing cost of products in the presence of complex numeric and categorical design attributes (cost drivers). The accuracy of the methodology presented in this work is assessed in comparison to a traditional approach, a polynomial regression model. The main concern behind this research is to predict the manufacturing cost of a product quickly and accurately without making assumptions on statistical distributions or estimating model parameters to simplify the complex relationship between categorical and numeric product design attributes and the manufacturing cost.