Vegetable commodity pricing and replenishment strategy based on GA-XGBOOST algorithm

In this paper, for the problem of vegetable commodity replenishment, sales volume and cost-plus pricing, firstly, we analyzed the distribution law and correlation relationship of sales volume of different categories and individual products of vegetables, then we used ARIMA time series model and wavelet analysis method to solve to get the relationship between sales volume change and time, and found that sales of flower and leaf category fluctuated more stably with the time series, and showed cyclic changes in the first two years; Then K-means clustering algorithm and Spearman correlation coefficient were used to derive the correlation of sales volume between categories and individual products. Through the K-means model, all individual products are divided into four categories, and the more groups, the greater the correlation. In this paper, support vector machine, random forest, linear regression, GA-XGBoost and other methods were used to establish the pricing and total sales volume regression model respectively, and GA-XGBoost was selected as the optimal model by combining the evaluation indexes. Finally, using the GA-XGBoost algorithm based on the sliding window tree, the time series data is converted into regression data to find the optimal replenishment and pricing prediction value of six categories of commodities for one week, and get the optimal value of daily revenue of the superstore. For example, if the sales unit price of the chili pepper category is set at RMB 20.642, the predicted sales quantity and superstore revenue will be 386.918kg and RMB 5296.856, respectively.