| A. Kanavos, G. Vonitsanos, Ph. Mylonas |
| Integrating Machine Learning Approaches for Consumer Behavior Analysis in Retail Transactions |
| 20th International Workshop on Semantic and Social Media Adaptation & Personalization (SMAP 2025), 27-28 November 2025, Mystras, Greece |
|
ABSTRACT
|
| Retail analytics has become essential for supermarkets, which generate vast volumes of transactional data on a daily basis. Harnessing this data to model consumer behavior enables informed decision-making in marketing, inventory management, and customer engagement. This paper proposes an integrated framework that combines association rule mining, k-means clustering, and Random Forest classification to extract and exploit insights from retail transactions. Association rules reveal frequent product co-occurrences, supporting interpretable strategies for cross-selling and shelf organization. Clustering uncovers distinct consumer segments, highlighting heterogeneity in purchasing patterns across shopper groups. Random Forest achieves superior predictive performance, demonstrating robustness in modeling transaction-level attributes and forecasting product-line preferences. Experiments on publicly available supermarket datasets validate the effectiveness of the framework, showing that specialized models outperform global baselines and that descriptive insights complement predictive results. By integrating descriptive, unsupervised, and supervised learning, the framework provides both actionable knowledge and interpretable forecasts, offering a comprehensive decision-support tool for modern retail environments.
|
| 27 November , 2025 |
| A. Kanavos, G. Vonitsanos, Ph. Mylonas, "Integrating Machine Learning Approaches for Consumer Behavior Analysis in Retail Transactions", 20th International Workshop on Semantic and Social Media Adaptation & Personalization (SMAP 2025), 27-28 November 2025, Mystras, Greece |
[  PDF] [
BibTex] [
Print] [
 Back] |
