Abstract:

Fashion color plays significant role in manufacture, marketing and profit for garment industries. Aiming at the problems existing in fashion color prediction studies on complicated color quantification, absence of time series parameters and establishment of prediction model performance relationship, influence of time series of defferent lengths on the the preformance of GM(1,1) fashion color hue prediction model was discussed, using Intercolor for womens's spring/summer as the subject, the PANTONE color system as the basis of dthe color quantification and mean absolute errors as the comprehensive evaluation index of the prediction effect. The result show that the grey GM (1,1) model established with 6-year time series has high overall precision on the prediction of fashion color hue, with the sum of mean absolute errors only of 1.9%.

Key words: fashion color, hue prediction, grey system theory, time series, mean absolute error