Clustering of Screw Press Machine Conditions using the K-Medoids Method

The screw press plays an important role in the oil extraction process; thus, monitoring its condition is essential to maintain performance and prevent failures. This study aims to cluster screw press machine conditions using the K-Medoids method. The dataset consisted of 23,002 records from PT. XYZ was collected in April–May 2024 with two attributes: temperature and pressure. The data was processed through selection, pre-processing, and transformation stages using z-score normalization before clustering. Model evaluation employed the Silhouette Coefficient and the Davies-Bouldin Index (DBI). The results show that the best configuration was at K = 7, with a Silhouette value of 0.5494 and a DBI of 0.5521, indicating a reasonable structure and good separation. Thus, the K-Medoids method has been proven effective in clustering screw press machine conditions and useful in supporting machine maintenance decision-making.