Implementation of K-Means Clustering Algorithm for Inventory Management Optimization at Putra Mart

Abstract

The rapid advancement of Information Technology has become a crucial element in enhancing business decision-making efficiency. Putra Mart, a prominent wholesaler in Bintuhan City, faces significant operational challenges due to its reliance on manual inventory recording and supply management systems. These manual processes often lead to data inaccuracies, stock imbalances, and difficulties in identifying market demand patterns. This research aims to address these issues by implementing data mining techniques using the K-Means Clustering algorithm to categorize inventory data into strategic groups based on stock levels and supply frequency.The system development follows the structured Waterfall model, which includes requirements analysis, system design using Data Flow Diagrams (DFD), coding with Visual Basic .Net 2010, and database management using SQL Server 2008. The clustering process utilizes the Euclidean Distance formula to measure the proximity between data points and centroids, effectively partitioning the items into two main clusters: "Fast Moving" and "Slow Moving" goods.The results of the analysis on 2020 transaction data successfully identified a clear distribution of products, with Cluster 1 accounting for 40% of the tested items. System testing through the Blackbox method confirmed that all functional features operate correctly, while a user satisfaction survey yielded a score of 84%, categorized as "Good." This study concludes that the implementation of the K-Means algorithm provides a reliable, data-driven solution for Putra Mart to optimize its inventory management, minimize deadstock, and improve overall service quality for its partner stores