<![CDATA[The annual admission of new students requires the early procurement of university jackets to ensure distribution during the inauguration ceremony. However, the lengthy production lead time necessitates ordering months before the actual sizing data is fully collected. This issue is further complicated by low participation rates in size registration and extreme population spikes. This study proposes a time series forecasting approach to predict the proportional distribution of jacket sizes rather than absolute quantities. Specifically, the research compares the performance of Holt’s Exponential Smoothing and the Weighted Moving Average (WMA) method using historical size proportions from 2019 to 2025. Walk-forward validation was employed to evaluate the models based on the Mean Absolute Percentage Error (MAPE). The results demonstrate that WMA outperforms Holt’s Exponential Smoothing by achieving a lower MAPE of 6.56%. By extrapolating the WMA proportions to the 2026 target of 5,250 students and mathematically integrating the 6.56% error rate as a safety stock buffer, the final procurement quantities for sizes S through XXXL were precisely determined. This proportional forecasting framework provides a robust, quantitative foundation for institutional supply chain management, allowing early and accurate ordering despite incomplete preliminary data.]]>
Copyrights © 2026
