<![CDATA[Oil losses in palm oil processing materials, such as Final Effluent, Empty Fruit Bunches, Kernels, Pressed Fiber, and Decanter Solids, pose significant challenges in ensuring production efficiency. FOSS-NIRS technology has been proven capable of quickly and efficiently detecting oil content, but its detection accuracy requires further analytical support. This study aims to develop a machine learning model that can accurately classify FOSS-NIRS data to detect oil losses that are either above the standard (red category) or below the standard (green category). By utilizing FOSS-NIRS data across five material categories, the proposed model is expected to provide precise predictions and support decision-making in palm oil production processes. The results of the study indicate that applying machine learning methods to FOSS-NIRS data can enhance the accuracy of oil loss classification, making it a potential solution for broader implementation in the palm oil processing industry to optimize production efficiency.]]>
Copyrights © 2025
