Leveraging machine learning in balancing inventories and analyzing market trends: a comprehensive systematic review

Rapid market changes, shorter product life cycles, and supply-chain disruptions have increased the complexity of inventory management and demand forecasting. This review examines how machine learning techniques and trend analysis improve forecasting accuracy, reduce uncertainty, and support intelligent inventory decision-making. From an initial pool of 116,377 studies, 78 papers were systematically selected for detailed analysis. The review covers supervised, unsupervised, reinforcement learning, deep learning, hybrid forecasting approaches, and emerging quantum machine learning techniques. Models such as Random Forest, XGBoost, LSTM, and Temporal Convolutional Networks are widely used for demand forecasting and inventory optimization, while reinforcement-learning methods support adaptive inventory control. The study also highlights recent advances in explainable AI, resilience-aware optimization, and sustainability-focused inventory systems, while discussing challenges related to data sparsity, explainability, and governance.