Kansei Clustering Using Design Structure Matrix and Graph Decomposition for Emotional Design

Conventionally, Kansei engineering relies heavily on the intuition of the person who uses the method in clustering the Kansei. As a result, the selection of Kansei adjectives may not be consistent with the consumer's opinions. Nevertheless, to obtain a consumer-consistent result, all of the collected Kansei adjectives (usually hundreds) might need to be evaluated by every survey participant, which is impractical in most design cases. Accordingly, a Kansei clustering method based on design structure matrix (DSM) and graph decomposition (GD) is proposed in this work. The method breaks the Kansei adjectives down into a number of subsets for the ease of management among the survey participants. In so doing, each participant deals with only a portion of the collected words and the subsets are integrated using a DSM-based algorithm for an overall Kansei clustering result. In order to differentiate the groups in the combined DSM further, graph decomposition (GD) is used to yield non-exclusive Kansei clusters. The hybrid approach, i.e., using DSM and GD, is able to handle the Kansei clustering problem. A case study on cordless battery drills is used to illustrate the proposed approach. The obtained results are compared and discussed.