<![CDATA[This study develops a web-based Smart Inventory system that integrates inventory recording, demand forecasting, inventory analytics, and replenishment recommendations for fabric stock control at BKR Textile Kudus. A Research and Development approach using the Prototype model was combined with quantitative comparative forecasting. The dataset comprised 400 fabric items and 12 months of stock-out history. Four one-month-ahead methods were tested: three-month Moving Average (MA), Weighted Moving Average (WMA) with 1:2:3 weights, Single Exponential Smoothing (SES) with alpha = 0.30, and rolling three-point Linear Regression (LR)-were evaluated through rolling-origin backtesting using MAD, MSE, and MAPE. Across 14,400 forecast-actual comparisons, MA produced the lowest average errors, with MAD of 5.167 rolls, MSE of 36.881, and MAPE of 5.166%. Inventory analysis identified three Critical items, 78 Low-stock items, 157 Normal items, and 162 Overstock items. The principal contribution is the integration of item-level comparative forecasting, safety stock, reorder points, inventory-status classification, and automated restock recommendations within one operational web platform. The system translates forecasting results into transparent decision information for a local textile company, although the findings remain specific to the one-year dataset and organizational setting examined.]]>
Copyrights © 2026
