<![CDATA[This study aims to address the uncertainty of raw material inventory in the furniture industry through the implementation of the Random Forest Regressor machine learning algorithm. The primary problem addressed is demand fluctuation, which frequently leads to stock management inefficiencies, including overstocking or material shortages that disrupt production processes. The research method employs a quantitative approach with an experimental design, developing a web-based system using the Flask framework and MySQL database. The data sample includes historical sales transaction records and Bill of Materials (BOM) data for furniture products, such as dining tables and minimalist chairs. Prior to modeling, the data underwent a preprocessing stage comprising data cleaning, handling missing values, and normalization to minimize the impact of noise on transaction data. Data collection was conducted through the extraction of internal databases, which were then processed through feature engineering stages based on temporal trends. The results demonstrate that the Random Forest model can predict future raw material requirements with high accuracy, evidenced by a coefficient of determination ($R^2$) of 0.84 and a Mean Absolute Error (MAE) of 5.4.5 These findings prove that a data-driven approach provides more precise stock requirement estimations than conventional methods. In conclusion, the integration of this predictive technology offers practical contributions to accelerating managerial decision-making and optimizing operational efficiency in the medium-scale manufacturing sector. The implications of this study support the theoretical development of artificial intelligence-based decision support systems in supply chain management.]]>
