Analysis of Garment Order Clustering Using the K-Means Clustering Algorithm

The garment industry generates diverse order data, requiring proper data management and analysis to support production planning and improve production efficiency. The main problem faced by Sumatera Konveksi is the suboptimal utilization of order data to identify order characteristics based on processing time, product type, order type, and total order quantity. This study aims to apply the K-Means Clustering algorithm to group garment order data based on similarities in their characteristics to support production decision-making. The method employed is K-Means Clustering, with the optimal number of clusters determined using the Elbow method and the clustering results evaluated using the Davies-Bouldin Index (DBI) and Silhouette Score. The results indicate that the optimal number of clusters is K=4. The clustering process converged at the eighth iteration, producing four clusters: C0 with 233 data points, C1 with 322 data points, C2 with 465 data points, and C3 with 200 data points. The evaluation results yielded a DBI value of 1.0888 and a Silhouette Score of 0.4949. These two metrics provide insights into the compactness of data within each cluster and the degree of separation between clusters. The resulting clusters are expected to serve as a basis for production planning, scheduling, and more effective resource management at Sumatera Konveksi.