HIERARCHICAL EVALUATION MODEL RESPONDING TO NONLINEARITY

概要

論文の詳細を見る
In this study, the hierarchical evaluation model responding to nonlinearity was constructed for the evaluation of a design solution. Considering visual information process in cognitive science and brain science, the evaluation model is a three-hierarchical structure consisting of three layers; design variable, element evaluation and image evaluation. Element evaluation and image evaluation were represented by the evaluation factor in the factor analysis. Moreover, the relations between the factors in each layer were modeled on the linear regression model. In cases it was difficult to be modeled on the linear model, they were modeled on a nonlinear neural network model. This hierarchical evaluation model was applied to wineglasses. As a result, the hierarchical evaluation model, which involved the linear model from the design variable to the element evaluation and involved the nonlinear model from the element evaluation to the image evaluation, was constructed. Consequently, the most appropriate model-application of each layer and the effect of minimizing nonlinearity were confirmed.
日本デザイン学会の論文
2002-03-31