Prediction of Water Vapor Permeability of Single Jersey Fabrics by Linear Regression and Artificial Neural Network Methods

Garments should maintain thermal balance of the body under various environmental conditions. In order to provide comfort for wearer, it is necessary for clothing to transmit water vapor from the body to the environment as fast as possible. In this study, water vapor permeability of single jersey knitted fabrics produced from different materials and different yarn and fabric properties were predicted with both linear regression and artificial neural network (ANN). Results showed that material, yarn count and fabric density, had an important effect on the water vapor permeability of the fabrics. Both linear regression and ANN were appropriate for predicting water vapor permeability of single jersey fabrics, but predicting performance of ANN was better. While R2 values of linear regression models were in the range of 0.70