Journal of Resource Management and Decision Engineering 5 Serial Number 16 2026 05 01 Simultaneous Use of Clustering, Classification, and Association Rules for Customer Churn Analysis in Retail Stores Simultaneous Use of Clustering, Classification, and Association Rules for Customer Churn Analysis in Retail Stores 1 21 EN Journal Article 2025 12 22 <p>Customer churn analysis plays a critical role in sustaining profitability in competitive retail environments. This study proposes an integrated data mining framework that combines clustering, classification, and association rule mining within an enhanced RFM-based model, referred to as RFFML, which extends the traditional Recency, Frequency, and Monetary indicators by incorporating Frequency Items and Length of Interaction variables. Using large-scale real transactional data from a retail store, customer segmentation is first performed through clustering to identify distinct behavioral profiles. Subsequently, several classification models are evaluated for churn prediction, with an ensemble-based Stacking classifier achieving the highest predictive performance. In parallel, association rule mining is employed to uncover frequent product combinations associated with loyal and churned customer segments. The main contribution of this work lies in the analytical integration of these complementary techniques within a single unified framework, enabling both accurate churn prediction and interpretable behavioral insights. Experimental results demonstrate that the proposed approach outperforms individual classifiers while providing actionable knowledge to support targeted retention strategies in the retail sector.</p> https://www.journalrmde.com/index.php/jrmde/article/download/303/384