A Clustering Approach to Affective Response Dimension Selection for Product Design

Abstract

Consumers’ affective responses (CARs) to the appearance of a product will greatly influence their purchasing decisions. In the product design field, researchers often provide adjectives so that consumers can express their subjective feelings. However, there exists both similarity and vagueness among these adjectives, which make it difficult to choose suitable adjectives for semantic differential (SD) experiments. This study proposes an approach based on factor analysis (FA), hierarchical clustering analysis (HCA) and k-means clustering. First, an SD experiments is conducted to collect CARs evaluated data about a series of selected product samples. Second, FA is applied to the CARs data to extract the factor loadings of the adjectives. Finally, the clustering approach, which comprises HCA and k-means clustering, is used to group these adjectives according to the factor loading results obtained from the FA. A case study of mobile phone design is used to demonstrate the effectiveness of the proposed approach. Three representative pairwise adjectives: coarse-delicate, unoriginal-creative and discordant-harmonious were obtained from the experimental results conducted on the initial set of 22 pairwise adjectives. This proposed method is very helpful to product designers during new product development. keywords:Factor analysis、Hierarchical cluster analysis、K-means clustering、Kansei engineering、Product design.