Predictive Intelligence for Pharmacy Inventory: Machine Learning-Based Medicine Demand and Stockout Forecasting

While excessive inventory can raise holding costs, storage expenses, and expiry-related waste, drug stockouts also impede timely access to necessary medications. Traditional pharmacy inventory management may not fully account for shifting demand patterns because it primarily depends on past sales, present stock levels, and human replenishment decisions. The promise of deep learning and machine learning for inventory planning and pharmaceutical demand forecasting has been demonstrated in recent studies [1], [2]. Nevertheless, stockout prediction and demand forecasting are frequently handled independently. This study presents MediSync, a predictive intelligence platform that combines stockout-risk prediction with pharmaceutical demand forecasting for pharmacy inventory management. Relevant features are created using historical sales, inventory levels, and temporal data. For demand forecasting, models such as Random Forest, XGBoost, Long Short-Term Memory (LSTM), and Linear Regression are taken into consideration and assessed using MAE, RMSE, and MAPE. Predicted demand, existing stock, and replenishment lead time are taken into account when identifying stockout risk, and accuracy, precision, recall, and F1-score are used to assess performance. Through data-driven inventory management, the framework seeks to decrease needless overstocking, enhance pharmaceutical availability, and facilitate proactive replenishment decisions.