Persona Creation With Fuzzy Cluster Method And Its Application

If we would change the situation from Made in China to Created in China, thekeys must be product design. In this competitive society, most companies are hungryfor a high efficient process of design. In the industrial design, such a hypothesis exists:if product could meet users’ goals, then they would feel satisfied and like to purchaseit. At the same time if we could realize it at a low cost, we must succeed in commerce.Based on this hypothesis, we can find the design criterion clearly: focus on the bestusers accurately; describe their goals and needs precisely; realize them at a low cost.To this criterion, the paper tries to introduce Persona model to industrial design tomake design process more efficient and systematic. Traditional Persona creation is aqualitative method through observations and interviews, which doesn’t perform wellin persuasiveness and repeatability. In this paper, I aim to explore a Persona creationbased on quantitative analysis, namely fuzzy cluster.First, I review literature from two aspects, applications and methods. I focus theresearch on industrial design. Compared with different methods, I propose fuzzycluster analysis. Second, Persona is explained from interactive design, industrialdesign, marketing theory and cognitive psychology, which establish Persona’stheoretical basis. Third, I introduce how to construct a quantitative Persona model indetail. At last I illustrate my idea through a sport earphone design.The core of quantitative Persona creation model is fuzzy c-means algorithm.This algorithm has two drawbacks: initial clustering prototypes are chosen randomly;the number of clustering classification c and weighting exponent m must be givenfirst. To the first problem, I build a potential function to define initial clusteringprototypes in the great density place. To the second, I try different value of c and m.When the result of fuzzy cluster show that the difference between classifications aresignificant and the similarity in the same classification is hign, this c and m areconsidered to be appropriate.