Application of Machine Learning in Predicting the Amount of Pharmaceutical Drugs Ordered for the Manufacturer

Predicting drug demand is a complex issue that has garnered significant attention from pharmaceutical manufac-turing and distribution companies. Recognizing the potential of machine learning algorithms in the healthcare supply chain, our research team embarked on utilizing these techniques to forecast the consumption of essential medications in Vietnam, particularly within the context of Nghe An province. We employed machine learning models such as Random Forest, LGBM, Histogram-Based Gradient Boosting, and XGBoost, focusing on the Nghe An province. By leveraging specific pharmaceutical transaction data in the region, we gained valuable insights into the dataset characteristics through data collection, preprocessing, and data analysis. We aimed to predict the usage demands for commonly used drugs by formulating a regression problem. Our initial results showed great promise, with the highest Root Mean Square Error (RMSE) of 0.95 and both R-squared and Adjusted R-squared values of 0.81 belonging to Random Forest. Through continuous endeavors, we aspire to enhance the accuracy and effectiveness of drug demand forecasting in the pharmaceutical sector and then optimize inventory management, improve resource allocation, and enhance service delivery to meet the healthcare needs of the people of Nghe An province and Vietnam as a whole.