<![CDATA[This study discusses the application of K-Means Clustering and Single Linkage algorithms in classifying the quality of Fresh Fruit Bunches (FFB) of oil palm based on four main variables: ripeness level, oil content, bunch weight, and water content. The dataset consists of ten harvest records representing variations in FFB characteristics. The analysis process began with Min–Max normalization, followed by clustering using K-Means (k=3) and Single Linkage (Hierarchical). The results showed that K-Means produced a lower Sum of Squared Error (SSE) (215.32), indicating more compact and homogeneous clusters, while Single Linkage achieved a higher Silhouette Score (0.914) compared to K-Means (0.892), reflecting clearer separation between clusters. This implies that K-Means is more suitable when compactness is prioritized, whereas Single Linkage performs better when the goal is natural separation. The implementation was carried out through a web-based system at PT. Teguh Karsa Wanalestari 2 Pom, Siak Regency, Riau, to address the problem of manual and subjective FFB quality assessment. The system enables real-time data input, automated calculations, evaluation with SSE and Silhouette Score, and presents results in tables, graphs, accuracy reports, and Gmail notifications. With this system, FFB quality can be consistently grouped into three grades (A, B, C), supporting decision-making in harvesting, distribution, and processing more objectively and efficiently to maintain Crude Palm Oil (CPO) quality.]]>
Copyrights © 2025
