Efficient Medical Supply Chain Forecasting Using Time Series Analysis

The unexpected shortage of potentially life-saving medication can result in the loss of lives, while overstocking can strain financial resources and generate medical waste. Due to this reason, forecasting techniques are of great concern to address this medical inventory management challenge. The paper presents an AI-based strategy for foreseeing medical store inventories using time-series analysis. The approach aims at predicting inventory variations and delivers order demand forecasts by incorporating temporal patterns and trends in historical data. Time series learning models are a subset of algorithms specifically developed to analyze and extract important insights from sequential data, where the order and timing of the data points are crucial. A comprehensive approach, namely the augmented dickey-fuller (ADF) test, is employed as a pre-processing technique to assess stationarity data and ensure statistical properties like mean and variance remain consistent over time. This work utilizes advanced time series forecasting methods and machine learning algorithms to model inventory fluctuations, capturing long-term dependencies and nonlinear patterns in the data. Out of all the applied techniques, seasonal autoregressive integrated moving average (SARIMA) outperformed autoregressive integrated moving average (ARIMA) as well as extreme gradient boosting (XGBoost). SARIMA showed minimum root mean square error which is 265.7 for order demand forecasting.