Abstract: Firstly, the factors influencing the sales volume of SKC (Stock Keeping Color) of retail products and their effectiveness are studied with SPSS and R softwares, combined with one-way ANOVA and deep regression analysis. Then, the sales volume of a target subcategory is predicted in grey model with R software, and the accuracy of the prediction data is obtained with Excel software combined with the mean absolute percentage error method. Finally, the prediction model is used to predict the sales and inventory of a target category. The results show that the sales method has a significant effect on the SKC sales volume of retail products; the factors influencing the SKC sales volume of retail products include inventory, discount and selling price; the average absolute percentage error of the target subcategory in each week is mostly less than or equal to 1, and the smallest is 0, which can be considered as a high prediction accuracy of the model. The prediction results of the future sales and inventory of a target subcategory show that the supply of this subcategory is obviously higher than the demand, and the inventory management strategy needs to be adjusted appropriately.

Key words: precise retail, demand prediction, one-way ANOVA, deep regression analysis, grey prediction