Objective evaluation model of visual elegance of swirl skirts - Part (II): neural network method

摘要

Using regression analysis, visual elegance of swirl skirts objective evaluation method was already established in Part I. Neural network is used to establish prediction model in this research to effectively evaluate visual elegance of swirl skirts. According to this research, the correlation coefficient between the visual elegance of swirl skirt practical and predicted value of model established by neural network method is 0.95. Neural network method is more accurate than regression analysis. Furthermore, it can establish visual elegance of swirl skirts prediction model with two key parameters of fabric mechanical properties and still preserves high fitting ability while keeping the benefits of simple, convenient to use, objective and effective.