A Data Mining Method for Fashion Design to Evaluate Consumers’ Emotional Preference

Abstract

The purpose is to propose a data mining method for fashion designers to build a representative design model reflecting consumers’ emotional preference. The present study consists of two modules: the extraction of Kansei information and the construction of the cluster model. A case study of men’s outer shirts is given to demonstrate the effectiveness of the proposed method. Based on the result of the cluster models in this study, four men’s outer shirts with different styles are produced as the representative design set. The selected design elements of each style represent the typical designs that satisfied consumers’ emotional preferences accordingly. A correlation between the Kansei intention of consumers and the discrete design elements can be established through the proposed method in order to predict the proper clothing style. The proposed method provides an effective means for fashion designers to identify the important design elements with positive impacts on specific consumers.