Predicting The Trends Of Home-textile Fashion Colors Based On Grey-markov Model And Support Vector Machine

Predicting the trends of Home-textile fashion colors is to judge and determine the color trendency of home-textile products in a period of time by the specific methods and models.Accurately grasp the trend of color can effectively assist the design work of home-textile products,avoid blind development of products,and thus enhance the competitiveness of enterprise products.In fact,the color creates low cost and high added value amazing competitiveness for products,which makes color's trendency information in a state of closure and confidentiality,thus has always caused that the fashion color information is not widely spread and apply in time,therefore,it promotes the prediction research of color trendency of home-textile products.At present,although some research achievements have been made in this field.However,overall,the study is still at groping stage,exists some problems in the quantitative classification of color data and low degree of data mining,it is necessary to establish and perfect the forecast system of home-textile fashion color.This subject took fashion color palettes as the data source,released by the international authoritative agency,analyzed the color scheme information and color system,confirmed the method of quantifying and classifying color data.Then,the prediction models of home-textile fashion color were constructed,which combined with the superiority of grey model,markov model,support vector machine and the characteristics of fashion color hue data.At last,a specific application study has been carried out on the prediction of home-textile fashion color hues.The content structure of this subject was arranged several parts as follows:Research background,significance and current status of relevant research work were discussed in chapter one,then research ideas and methods of the project were presented.Part two expounded relevant knowledge of the color system,and opted for the method of quantifying and classifying color data according to the information of fashion color palettes.Finally,characteristic analysis of the quantized hue data was carried out.The basic theory of data prediction was introduced in chapter three.Based on this,it discussed the practical feasibility of several common prediction models applied for the trend prediction research of home-textile fashion color.Part four deeply studied the feasibility of applying grey model to the prediction of home-textile fashion color hues,then designed concrete prediction scheme and implementation steps.On this basis,the improved method based on markov residual correction was proposed,and both of them on prediction effect were compared.Part five dug into the feasibility of using support vector machine to the prediction of home-textile fashion color hues,established in support vector machine which has strong learning ability in nonlinear relationship and good advantage on small sample date modeling,designed specific prediction scheme and implementation steps of applying support vector regression model to the prediction of home-textile fashion color hues.At last,different model parameters have influences on prediction results and model's generalization ability by constantly adjusted and optimized model parameters were contrasted.Last part is the conclusion of this paper.It summarized the study conclusions of subject,and offered some questions and directions to be considered in the future research work.