Predictive Analytics in Retail Pharmacy: QR Code Implementation on Purchase Intentions Using Random Forest

In this study, we explore the efficacy of a Random Forest model in predicting customer repurchase intentions in a retail pharmacy setting that has adopted QR code technology. The model achieved a training R² of 0.6964 and a testing R² of 0.4558, indicating that while it captures a substantial proportion of the variation in the training data, its performance on independent data calls for cautious interpretation. Notably, feature importance analysis highlighted that store attributes and service quality are critical determinants of repurchase behaviour, explaining 32.91% and 29.86% of the variance, respectively. Additional factors, such as the physical store environment and waiting time, also played significant roles. These findings suggest that both tangible store features and digital interventions contribute to customer loyalty. The integration of rigorous model validation techniques, including cross-validation and hyperparameter tuning, reinforces the model’s potential as a robust predictive tool. This work provides meaningful insights for retail pharmacies aiming to optimize customer engagement through a blended approach of digital innovation and traditional service excellence.