APPLICATION OF DATA MINING TO PREDICT DISTRO CLOTHING SALES USING THE K-MEANS CLUSTERING METHOD

Shafa Arrizqa Az Zahroh (University of Muhammadiyah Sidoarjo, Indonesia)
Nuril Lutvi Azizah (University of Muhammadiyah Sidoarjo, Indonesia)
Novia Ariyanti (University of Muhammadiyah Sidoarjo, Indonesia)
Irwan Alnarus Kautsar (University of Muhammadiyah Sidoarjo, Indonesia)

Journal of Artificial Intelligence and Digital Economy
Vol. 2 No. 12 (2025): Journal of Artificial Intelligence and Digital Economy

Objective: The research aims to classify distro clothing products at Aldi Store according to sales levels to support more effective inventory and marketing strategies. Method: Data processing was conducted using Google Colaboratory, applying the K-Means Clustering algorithm combined with evaluation metrics including the Silhouette Coefficient, Calinski-Harabasz Index, and Davies-Bouldin Index to determine the optimal cluster structure. Results: The analysis shows that K-Means successfully groups sales patterns with strong cluster performance, indicated by a Silhouette Coefficient of 0.576, a Calinski-Harabasz Index of 19.125, and a Davies-Bouldin Index of 0.308, reflecting high cohesion and clear separation among clusters. Novelty: This study integrates multiple validity indices within a practical retail context, demonstrating a robust clustering approach that enhances customer segmentation accuracy and provides actionable insights for strategic decision-making in product management.