<![CDATA[The rapid growth of transactional data in the retail sector, particularly in supermarkets, has resulted in large and complex sales datasets that require effective analytical methods to identify meaningful sales patterns. This study applies data mining through clustering techniques by implementing the K-Means algorithm using Python to classify supermarket sales patterns and group transactions with similar characteristics. The research methodology includes data collection, preprocessing, normalization using StandardScaler, determination of the optimal number of clusters through the Elbow Method, clustering with the K-Means algorithm, and evaluation of clustering quality using the Silhouette Score, Davies–Bouldin Index, and Inertia (Within-Cluster Sum of Squares). The results indicate that from 1,000 sales transactions, three clusters were formed, consisting of 515 transactions in the low sales cluster, 314 in the medium sales cluster, and 171 in the high sales cluster. The clustering evaluation produced a Silhouette Score of 0.6055, a Davies–Bouldin Index of 0.502, and an Inertia value of 122.01, indicating that the resulting clusters are compact and well separated. These findings confirm that the K-Means algorithm is effective for grouping supermarket sales patterns and can support sales management and strategic planning.]]>
Copyrights © 2026
