Data Mining Model And Application For Affective Design

Affective design has received much attention from both academia and industries.It aims at incorporating customers’ affective needs into design elements that delivercustomers’ affective satisfaction. The main challenge for affective design originatesfrom difficulties in mapping customers’ subjective impressions, namely Kansei, toperceptual design elements. This paper intends to develop an explicit decision supportto improve the Kansei mapping process by reusing knowledge from sales records. Asone of the important applications of data mining, association rule mining lends itselfto the discovery of useful patterns associated with the mapping of affective needs. AKansei mining system is developed to utilize valuable affect information latent incustomers’ impressions of existing affective designs. The inference utility of each ruleis evaluated according to its achievements of customers’ expectations. Conjointanalysis is applied to measure the expected and achieved utilities of a Kansei mappingrelationship. Then, both utilities are normalized. Based on inference utilities, mappingrules are further refined to empower the system with useful inference patterns. Thesystem architecture and implementation issues are discussed in detail. An applicationof Kansei mining to affective design of Volve long haul trucks is presented.