Fashion retailing forecasting based on extreme learning machine with adaptive metrics of inputs

Abstract

In the fashion retail industry, a versatile sales forecasting system is crucial for fashion retailers. In order to avoid stock-out and maintain a high inventory fill rate, fashion retailers require specific and accurate sales forecasting systems. In this study, a hybrid method based on extreme learning machine model with the adaptive metrics of inputs is proposed for improving sales forecasting accuracy. The adaptive metrics of inputs can solve the problems of amplitude changing and trend determination, and reduce the effect of the overfitting of networks. The proposed algorithms are validated using real POS data of three fashion retailers selling high-ended, medium and basic fashion items in Hong Kong. It was found that the proposed model is practical for fashion retail sales forecasting and outperforms the auto-regression (AR), artificial neural network (ANN), and extreme learning machine (ELM) models.