<![CDATA[Abstract: This study aims to analyze and classify the level of drugdemand and provide drug procurement recommendations at UPT Puskesmas Air Tiris using the K-Means Clustering method. The main problem identified is that the drug demand analysisprocess is still conducted through data recapitulation, making it difficult to provide optimal information regarding drug demandlevels. The data used in this study were obtained from the DrugUsage and Request Report (LPLPO) for the period from January 2025 to January 2026, consisting of initial stock, receipts, inventory, usage, and remaining stock variables. The research methodology applies the Knowledge Discovery in Database (KDD) process, including data selection, preprocessing, transformation, data mining, and evaluation. The clustering process uses three clusters representing high, medium, and low demand levels. The results show that 309 drugrecords were successfully grouped into three clusters, consistingof 4 drugs in Cluster 1, 27 drugs in Cluster 2, and 278 drugs in Cluster 3. The clustering results were subsequently used as a basis for procurement recommendations by considering the average monthly demand, remaining stock, and a safety stock of 10%. The results indicate that the implementation of K-Meanscan assist in identifying drug demand levels and provideprocurement recommendation information as a reference for improving drug inventory management at Puskesmas Air Tiris.]]>
Copyrights © 2026
