Implementasi Algoritme K-Means++ Untuk Clustering Penjualan Bahan Bangunan

Abstract

Utilization of the K-Means++ Clustering algorithm in the sales of building materials clustering system at UD Sumber Bangunan. Currently, the store does not use computers to run its system, resulting in transaction data being used merely as archives without being optimally utilized for marketing strategies and business decision-making. This research aims to examine whether the use of the K-Means++ Clustering algorithm can provide advantages in forming better and more efficient clusters for building material sales data. Evaluation is conducted by comparing K-Means++ with K-Means using the evaluation metrics DBI (Davies-Bouldin Index) and Silhouette Coefficient. The evaluation results show that K-Means++ outperforms using K-Means alone. The DBI values are lower for K-Means++, and the Silhouette Coefficient is higher for K-Means++, indicating that K-Means++ produces better-defined clustering. The utilization of the K-Means++ Clustering algorithm provides benefits in business decision-making at UD Sumber Bangunan, assisting in reducing stockpiles and enhancing customer satisfaction. Additionally, the clustering system built using the Waterfall method also contributes positively to achieving the set objectives..