Kansei Design Methodology Research And Its Application Based On Personal Construct Theory And Variable Precision Rough Sets

The measurement of users’ Kansei,the definition of design elements through users’ Kansei and the extraction of Kansei knowledge are three crucial issues that exist in the methodology research of Kansei engineering.In this article Kansei knowledge is defined as a logic knowledge structure that reveals the association mechanism between users’ Kansei and the design elements of specified product.Kansei engineering is a kind of design optimization technique focusing on users’Kansei needs.Nevertheless,Kansei needs is essentially vague,uncertain,mixed with linear and nonlinear relationship and abnormal distribution.Today,systematic and normalized exploration on the three crucial issues that considering the complicated specialty of users’ Kansei needs is lacking.In this article,Personal Construct Theory(PCT)is introduced in the process of Kansei measurement and the definition of design elements.Meanwhile,Variable Precision Rough Set(VPRS)is introduced to acquire Kansei knowledge.Through these endeavor,a new Kansei design methodology is developed considering the complicated specialty of users’ Kansei needs.The main research contents of this article are as below.(1)The generation method of initial Kansei image words set based on Repertory Grid Method(RGM).Firstly,RGM is applied to the export of target users’ personal constructs.Then,constructs are transferred to corresponding Kansei image word set on the guidance of laddering.(2)The extraction method of representative Kansei image word pairs based on Generalized Procrustes Analysis(GPA).Firstly,the initial Kansei image evaluation data is defined as Free Choice Profiling(FCP)dataset.Then,GPA is adopted to extract consensus grid that could reflect general tendency of the evaluation data.After that,Principle Component Analysis(PCA)is applied to map the initial Kansei word to the data space of consensus grid,and the structural coefficient of each initial Kansei word on every principle component is calculated.At last,the representative Kansei image word pairs are extracted from the initial Kansei words set considering the EPA semantic space and other factors.(3)The definition of critical design elements that affect users’ perceptual cognizance based on Pyramiding and the refined analysis method of scan path.(4)The extraction method of Kansei association rules based on VPRS.Firstly,fundamental concepts like upper appropriation and so on are defined considering case of threshold value β.Then,the solving algorithm of optimal β reduction which take mutual information gain as the heuristic information is proposed.At last,the algorithm of generating Kansei association rules is proposed based of the concept of rule’s support degree and confidence degree.(5)The implementation process and the existing problems of the proposed Kansei design methodology are presented explicitly in the application study of hair dryer form design.And eventually,an example of extracting Kansei association rules from decision table which contains tow decision class is presented considering target Kansei fashion-not fashion.The achievement of this article is an important expansion of Kansei design theory which provides a new way and methodology foundation for the selection of representative Kansei image word pairs and the acquisition of Kansei knowledge.Meanwhile,the output Kansei association rules could connect to the inference engine of Kansei engineering system smoothly which roundly consider the complicated character of users’Kansei needs.So,the proposed methodology possesses favorable application value.