Kansei evaluation and clustering in emotional design

Many products launched in the market failed despite having excellent functionality because they failed to appeal to consumers through their design. Design speaks a thousand words to the consumers and every design conveys a different message. Globalisation brings about an increasing number of affluent societies and more people want a product that suits their individual personality such as professional or sporty. It is important to know what emotional impacts or Kansei the consumers want in a product before designing so as to convey that message of appeal. The purpose of this study is to help designers or researchers understand consumers and their emotional needs. The aim is to create a tool useful for them to generate Kansei clusters for analysis of all the emotions consumers experienced to a particular product and eventually know how to design for their targeted group of consumers. Extensive literature reviews of current tools and methodologies, not limited to the area of Kansei Engineering, was done to investigate and propose an improve method to facilitate future applications in Kansei Engineering. The author modified an existing method, used generally in Kansei Engineering to cluster Kansei words. On top of that, the author incorporated two other analytical models, the Principal Component Analysis and Cluster Analysis, to obtain a more objective result. The existing method poses a limitation in Kansei Engineering because designers or researchers have to assume a threshold value, based on their gut’s feeling, to determine the final number of Kansei clusters. The result may not be ideal and may be influenced by the designer’s emotions. Hence, the author implemented a new selection criterion α to determine the optimal number of clusters that based objectively on consumers’ emotions after a dendrogram was produced through hierarchical clustering. A case study was done to validate the proposed methodology and a fair comparison of the results with an existing method was made to evaluate the robustness of the proposed methodology. 10 Kansei tags were produced for 32 Kansei words as opposed to only 8 Kansei tags using the existing method. The disparity was mainly due to the different methods used to determine the final number of clusters. A further analysis of the proposed selection criterion α was done to assess the soundness of the equation.Bachelor of Engineering (Mechanical Engineering