Fashion colour forecasting based on BP neural network

Xiaohong Kong¹, Lixia Chang², Junpeng Xu¹

¹School of Mechanical and Electrical Engineering, Henna Institute of Science and Technology, China, 453003, Henan, China

²School of Fashion, Henna Institute of Science and Technology, China, 453003

Fashion colour forecasting is a hot issue in fashion industry and also a hard problem because of much uncertain information. Utilizing strong mapping capability of BP Neural Network (BPNN) for nonlinear function, this paper investigated the forecasting model of fashion colour. Based on colour data of recent several years, the forecasting model for future colour trend is discussed and built. The historical data are input to train the Neural Network weights and the different parameters of BPNN were investigated to find how to affect the forecasting performance. The results demonstrate the algorithm is very efficient in colour forecasting and can approximate nonlinear relationship of fashion colour very close.