Regression Analysis for Predicting the Hemline of a Statically Draped Shape

概要

論文の詳細を見る
We are developing a system within the limitations of a statically draped shape which presents many shapes satisfying the reference conditions, such as the number of nodes, and simultaneously indicates the mechanical properties. In this system, it is required to present many draped shapes corresponding adequately to the given reference conditions. For this reason, we have tried to consolidate the various draped shapes calculated in a garment simulation as a regression equation.<BR>As the first step, we have tried to made regression equations of a hemline shape. The explanatory variables are the number of nodes, existence/absence of contact, bending length, pendant length, and radius of a support stand. Although the response variables are coordinate values of the hemline, they are changed to Fourier coefficients selected to reduce the number of variables. The number of selected Fourier coefficients, i.e., the number of regression equations, was a total of 220. Two hemlines, one of which calculated using the regression equation and another determined using the original garment simulation, had a relationship with a high coefficient of determination and were well in agreement.<BR>By using the regression equations, it was possible to design an arbitrary hemline shape corresponding to the conditions of the number of nodes and existence of contact.
社団法人日本繊維製品消費科学会の論文