Performances of Artificial Intelligence Hybrid Models' in Prediction of Clothing Comfort From Fabric Physical Properties

Abstract

The purpose of this paper is to compare different approaches toward the prediction of human psychological perception of clothing comfort performance from fabric physical properties. A series of running wear trial, which involved 8 sets of tight-fit garment and 38 subjects, was conducted in an environmentally controlled chamber. Thirty-three fabric physical property indexes were measured and 9 individual sensations (clammy, sticky, breathable, damp, heavy, prickly, scratchy, tight and cool) and overall comfort were rated by the subjects during the running period. Different predictive models were derived on the basis of human perception process to predict clothing comfort from fabric physical properties. Results show that hybrid models consist of data reduction (abstract fabric properties and sensations into physical and sensory factors respectively), self-learning capability (to learning the behaviour of individuals) and fuzzy reasoning (handle the fuzziness between sensory factors and overall clothing comfort) able to generate better predicted clothing comfort score than other predictive models in this study The correlation between predicted and actual clothing comfort rating was 0.739, with significant level of 0.001.