Optimization of K-Means Clustering Method by Using Elbow Method in Predicting Blood Requirement of Pelamonia Hospital Makassar

Abstract

Hospitals require an adequate supply of blood to meet patient needs. Accurate prediction of blood demand is essential to optimize inventory management and avoid shortages or overstocks. This study aims to predict blood demand at Pelamonia Hospital using K-Means Clustering and Elbow methods. Historical data on blood demand at Pelamonia Hospital was collected and processed. The Elbow method is used to determine the optimal number of clusters in the K-Means Clustering algorithm. Sum of Squared Errors (SSE) or Within-Cluster Sum of Squares (WCSS) values were calculated for various clusters, and the elbow point on the graph of SSE/WCSS vs. number of clusters was identified as the optimal number of clusters. Once the optimal number of clusters is determined, the K-Means Clustering algorithm is applied to the blood demand data, resulting in grouping the data into specific clusters. Each cluster is analyzed to find interesting patterns or characteristics, such as clusters with high or low blood demand. From the results of the SSE calculation process on 1057 blood demand data, the result that has the biggest decrease is at k = 4 with a difference value of 2754.90. The clustering results and patterns found are used to predict future blood demand by identifying which cluster best fits the current or expected conditions. The characteristics of the clusters are used to estimate the likely blood demand. This approach provides valuable insights into blood demand patterns and enables hospitals to better anticipate blood demand, thereby optimizing inventory management and improving the quality of healthcare services.