G-LMBPNN: A New Fashion Color Prediction Model

Since the current fashion color forecasts have some disadvantages in practical application, there is considerable interest in building models that can predict fashion value of the colors precisely and swiftly from historical data. This paper proposed a new forecasting model called G-LMBPNN (Gray Levenberg-Marquardt Back Propagation Neural Network). It utilizes gray process to obscure the data sequence and learns the nonlinear relation through optimized BP neural network training. Finally, we whiten the simulation sequence to get the predicted value. We show the effectiveness of G-LMBPNN through a comprehensive experimental evaluation based on three models.