A neural network approach to eco-product form design

This paper presents a neural network (NN) approach for determining the best design combination of product form elements that match a given product value represented by eco-product value (EpV) attributes. Twenty-seven representative office chairs are derived from 100 collected as the experimental samples by using multidimensional scaling and cluster analysis. Moreover, a morphological analysis is applied to extract 7 product form elements from these sample office chairs. The concept of Kansei Engineering and a best-performing NN model are chosen to examine the complex relationship between 7 design elements and 15 consumers' perceptions of EpV attributes which are identified and categorized into aesthetic, functional, and environmental dimensions. With the NN model, an office chair design database is built consisting of 960 different combinations of design elements, together with their associated EpV attributes. The application of the design database provides product designers with the best combination of product form elements for examining the aesthetic, functional, and environmental-friendly attributes to an office chair design, and facilitating the eco-product form deign process.