EXPLAINABLE MACHINE LEARNING FOR DYNAMIC PRICING IN FAST-CHANGING RETAIL ENVIRONMENTS

Abstract

Dynamic pricing has become a critical strategy in fast-changing retail environments where demand patterns, customer behavior, and competitive conditions evolve rapidly. While machine learning models have shown strong performance in optimizing prices, their lack of transparency limits trust, adoption, and regulatory compliance in real-world retail applications. This paper proposes an explainable machine learning framework for dynamic pricing that balances predictive accuracy with interpretability. The approach integrates advanced pricing models with explainability techniques to reveal how key factors such as demand fluctuations, inventory levels, customer sensitivity, and competitor pricing influence price decisions. By providing human-understandable explanations alongside price recommendations, the framework enables retailers to make informed, accountable, and adaptive pricing decisions in real time. The model is designed to operate effectively in volatile retail settings, including online marketplaces and omnichannel platforms, where rapid changes require continuous learning and justification of pricing actions. Experimental evaluations demonstrate that the proposed explainable approach achieves competitive revenue performance while significantly improving transparency and decision confidence compared to black-box models. The findings highlight the importance of explainable machine learning in bridging the gap between algorithmic efficiency and managerial trust, supporting ethical, compliant, and sustainable dynamic pricing practices in modern retail ecosystems.