<![CDATA[This study aims to develop an integrative model based on data mining and machine learning to analyze purchasing patterns and predict consumer product sales at CV. Mitra Ridge. The approach used combines the FP-Growth algorithm to discover product association patterns (frequent itemsets and association rules), XGBoost as a gradient boosting-based sales prediction model, and SHAP (SHapley Additive Explanations) to provide transparent interpretability of the model's prediction results. The data used is sales transaction data from October 2024 to September 2025, which includes 500 transactions with various types of consumer products. The results show that the integration of association pattern features from FP-Growth as additional input to the XGBoost model can improve prediction accuracy compared to a single XGBoost model without integration. SHAP analysis revealed that purchase frequency, product category, and product combination patterns are the most influential factors in sales prediction. This integrative model is proven to be superior in performance and provides deeper business insights, so it can be used as an operational decision support system in stock management, bundling strategies, and data-driven marketing planning in consumer retail companies. ]]>
Copyrights © 2026
