<![CDATA[Uncertainty in market demand poses a fundamental challenge in e-commerce supply chain management. This study evaluates the accuracy of daily sales forecasting for the "Set" product category in the Amazon Sales Report dataset by comparing the traditional ARIMA model with the modern additive Facebook Prophet model. Inventory management in e-commerce is often hindered by unpredictable demand fluctuations, which are difficult to forecast manually. The findings reveal that Prophet outperforms ARIMA, achieving a mean absolute error (MAE) of 35.412 and a root mean square error (RMSE) of 48.723, corresponding to an 18.82% improvement in forecasting efficiency. Prophet’s ability to capture weekly seasonal patterns demonstrates its suitability as a more reliable approach for operational stock management.]]>
Copyrights © 2026
