Quantitative research on product form based on Kansei engineering

Abstract

In mature market competition, the evaluations of users on product design are significantly dependent on the perception of product forms. In order to study the mapping relationship between product form and perceptual image, a quantification algorithm model was proposed. Initially, samples and perceptual image word pairs were collected and selected on basis of helmet products. A score was given to each sample image by subjects, and the initial data were obtained. Subsequently, the data were pre-processed using the Leida criterion for rejecting the abnormal perceptual image. The samples were grouped into 12 categories by clustering analysis. Finally, a multidimensional scaling and linear regression analysis were utilized to generate the joint space map of perceptual attribute vectors and sample scattered points. The attribute value of the sample was obtained by calculating the distances from the scattered points to the projection points of the perceptual vector and the origin. Thereafter, the correspondence between the sample and perceptual words was obtained. Results demonstrate that the perceptual characteristics of words are more evident for the absolute attribute value larger than 2, rather than the absolute value less than 0.1. The larger absolute attribute value shows that the sample form has more perceptual characteristics of the vector. The study provides an effective method for quantifying the perceptual characteristics of product form.