Automatic Pricing and Replenishment Decision-Making for Vegetable Commodities Based on Seasonal Index Prediction Models

Abstract

This study provides an in-depth analysis of the relationship between sales, pricing, and product categories to construct a predictive model designed to optimize replenishment decisions and maximize superstore profits. The study begins with data preprocessing, integrating data coded according to individual products, on which the model is built and solved. In the first part, descriptive statistical analysis and visualization revealed that the sales volume of each category showed a cluster-like distribution, which was affected by seasonal variations. Based on the time series smoothness test, it is initially judged that each individual product is a smooth series, and the smoothness of the series is verified by ADF test. In the second part, the correlation coefficients of the sales volume of each category and single product were calculated by using Spearman correlation analysis, and the results were presented through heat maps, which revealed the strong correlation between different categories. For the relational equation between total sales volume and cost-plus pricing, an autoregressive and least squares model was developed, with cost-plus pricing as the dependent variable and total sales volume as the independent variable. For the seasonal index prediction model, the total daily replenishment and pricing at future dates were predicted by calculating the seasonal index. The results show that the replenishment volume of flowers and leaves fluctuates greatly, and it is recommended that pricing and replenishment volume be adjusted in a timely manner according to the actual sales situation; cauliflower, aquatic roots and tubers, and eggplant are relatively stable, and they should be adjusted moderately according to the market demand; and chili peppers and edibles are on a downward trend, and it is recommended that replenishment volume be reduced in order to avoid the inventory backlog.