Mining Customer preferences with ANN-DT

In the early stage of product design, mining customer preference is the key issue to the development of satisfactory products. This paper builds a preference model from a set of historical transaction to catch what customer wants in mass customization. An artificial neural-network decision tree algorithm (ANN-DT) is employed to model customer preference on different product families. It is able to extract with rules of customer preference with high fidelity. To illustrate the effectiveness of the proposed method, a case study of personal computers is reported.