Implementasi Data Mining dengan K-Means Clustering untuk Memprediksi Pengadaan Obat

Abstract

Community Health Center is one of the institutions that provides healthcare services. To ensure the provision of quality healthcare services, the Community Health Center management must be able to effectively manage medicine inventory to avoid the risks of shortages or excess stock. Therefore, the purpose of this research is to observe and perform clustering of medicine demands at Puskesmas Mandala using the K-Means Clustering technique. The data used includes medicine demand data from January to December 2023 at the health center. In its implementation, the RapidMiner application or software is utilized to perform clustering using the K-Means Clustering algorithm. The available medicine data will be grouped into 3 clusters: cluster 0 for high medicine demands, cluster 1 for moderate medicine demands, and cluster 2 for low medicine demands. Out of the 28 test data used, the results show the first cluster consisting of 24 items, the second cluster consisting of 3 items, and the third cluster consisting of 1 item with a Davies Bouldin Index value of 0.276. From this research, the Puskesmas can continue to procure medicine for the types classified under high-demand clusters to ensure that the medicine needs are consistently met.