Research on Pricing and Restocking Strategies Based on Hierarchical Clustering and ARIMA Model

This study aims to explore pricing and restocking strategies based on hierarchical clustering and ARIMA modeling. After conducting in-depth analysis of supermarket vegetable sales data, we initially preprocessed the data, including data merging, categorization, and calculation of sales volumes. Utilizing hierarchical clustering analysis, we categorized products into three groups, revealing the correlation between the sales volumes of different products. Correlation analysis indicated that there is no significant correlation between vegetable sales volumes, suggesting insensitivity to price. We further applied the ARIMA model for time series forecasting of vegetable sales volumes to predict the upper limits of sales for the upcoming week. The parameters of the ARIMA model were determined through auto-correlation and partial auto-correlation analysis, and the model's effectiveness was validated through model fitting and residual white noise testing.