Applying Artificial Neural Networks to Total Hand Evaluation of Disposable Diapers

Abstract

The objective hand measurement and also handle of the materials used for disposable diapers has been looked into, with consideration given the aspects of both dermatitis and comfort. In this work, we tried to predict the handle of disposable diapers by their physical properties using a back-propagation network and a stepwise regression. Handle properties of diapers were measured by universal test equipment and hand values of the fabrics were determined by a group of panelists consisting of some textile experts. The optimum construction of neural network was investigated through the change of layer and neuron number. The results showed that the back-propagation network could predict the hand values of disposable diapers with a meaningful difference. These disposable diapers were used to show that the results of neural network were in good agreement with subjective test results.