Features selection of stonewashed denim to a sensory descriptor

Abstract

A novel approach of physical feature selection is proposed based on the techniques of fuzzy logic and ordered weighted averaging (OWA) operators. The selection is considered as a Multi Expert-Multi Criteria Decision Making (ME-MCDM) problem, since this approach ranks physical features according to their relevancy to multiple experts’ perception. In this study, objective and subjective evaluations of stonewashed denim fabric were made to predict fabric handle in a simple way. The aim of this research was to select the most relevant objective parameters predicting a given subjective one with a minimum number of laboratory measurements. For this purpose, a wide database of treated denim fabrics was prepared by making subjective and objective evaluations. Subjective assessment was evaluated by two different panels. The first one consists of sensorial experts who classify the fabric samples according to their sensitivity to a predefined sensory descriptor. The second one is a panel that masters professional knowledge only in mechanical features of fabrics. The objective measurements were evaluated mainly by means of the Kawabata Evaluating system for fabrics (KES-F), Fabric Assurance by Simple Testing (FAST) system and Universal Surface Tester (UST). The soft computing modelling demonstrate that the fuzzy sensitivity and OWA operators are capable of selecting the most relevant objective parameters notably by introducing a smart percolation method yielding automatically the set of the most objective parameters.