A Kansei evaluation approach based on the technique of computing with words

• This paper presents a Kansei evaluation approach based on the technique of computing with words.
• Kansei preferences are modeled by positively worded items with 7 levels of semantic labels.
• Fuzzy relation-based clustering is used to extract a set of Kansei attributes from collected Kansei words.
• A cluster validation index is proposed to assist evaluators in determining the best number of clusters.
• Linguistic aggregation is used to synthesize Kansei priority information and rank the order of product alternatives.

Kansei evaluation plays a vital role in the implementation of Kansei engineering; however, it is difficult to quantitatively evaluate customer preferences of a product’s Kansei attributes as such preferences involve human perceptual interpretation with certain subjectivity, uncertainty, and imprecision. An effective Kansei evaluation requires justifying the classification of Kansei attributes extracted from a set of collected Kansei words, establishing priorities for customer preferences of product alternatives with respect to each attribute, and synthesizing the priorities for the evaluated alternatives. Moreover, psychometric Kansei evaluation systems essentially require dealing with Kansei words. This paper presents a Kansei evaluation approach based on the technique of computing with words (CWW). The aims of this study were (1) to classify collected Kansei words into a set of Kansei attributes by using cluster analysis based on fuzzy relations; (2) to model Kansei preferences based on semantic labels for the priority analysis; and (3) to synthesize priority information and rank the order of decision alternatives by means of the linguistic aggregation operation. An empirical study is presented to demonstrate the implementation process and applicability of the proposed Kansei evaluation approach. The theoretical and practical implications of the proposed approach are also discussed.