Optimasi Manajemen Stok melalui Pengelompokan Produk Ritel Berdasarkan Data Penjualan dengan K-Means

Gumilang, Mauly Chandra

Jurnal Algoritma
Vol 22 No 2 (2025): Jurnal Algoritma

This study aims to categorize products based on sales data to improve decision-making efficiency in traditional retail stores. The main problem is the suboptimal use of transaction data in storage and promotion strategies. The method used follows the CRISP-DM stages, with the K-Means algorithm applied to the attributes of sales volume, cost price, selling price, and total sales. The data analyzed included 400 transactions from March to October 2024, involving 175 products. The determination of the optimal number of clusters was carried out using the Elbow and Davies-Bouldin Index methods, which showed the best results at K=2. Evaluation using the Silhouette Coefficient resulted in a value of 0.882, which indicates good cluster quality. The products were divided into two categories, namely fast-moving and slow-moving items. These findings prove the effectiveness of K-Means in supporting data-driven stock and promotion strategies.