In order to satisfy the consumers’ pursuit of diversification of goods, new retail enterprises begin to gradually produce small quantities and various kinds of products, which makes the sales data become more complex and various, and then makes the inventory management more difficult. Therefore, it is very necessary to establish an accurate demand prediction model for the sub-category stratum. In this paper, we firstly consider the effect of external macrofactors on sales, and establish a multiple linear regression model to forecast the sales of the target products. Then we consider the regularity and trendency of previous sales, comparing the fitting degree of different parameter ARIMA models, and finally establish the ARIMA (2, 2, 1) model with the best prediction effect. Finally, in the light of the fitting degree, the two models are given different weights, and a predictive model that combines multiple linear regression and ARIMA (2, 2, 1) is established. It can be shown from the results that the prediction effect of combined model is better and it can accurately predict needs for new retail goods, thereby reducing the difficulty of inventory management and improving corporate competitiveness.
Loading....