Predicting Garment Appearance Quality

Developing steps in designing a system for predicting garment appearance quality are presented. The articles of clothing investigated are ladies’ jackets and the accent is given to subjective evaluation of individual factors of garment appearance quality. A group of semi-skilled evaluators and a group of experts in the field of garment engineering were doing the evaluation. This system of dual evaluation and comparison of the grades obtained are key factors in the development of the predicting models using the method of machine learning. The algorithms of the latter are implemented in the Orange machine learning system, which was used to develop the system for predicting garment appearance quality grade. It is founded on previous knowledge of correlations among the quality grade of a particular factor of garment appearance quality and the parameters of mechanical and physical properties of the analysed fabrics, incorporated into the garments in question. The workings of the system designed result in a prediction of garment appearance quality grade, presented separately for each individual factor of appearance quality defined.