Demand Forecasting in Supply Chain Using Artificial Intelligence Basis Machine Learning

There is a lot of rivalry among stores in this day and age of technology. Predicting future demand or sales is one of their biggest obstacles. The aim of this work is to forecast department-wide sales for Walmart stores using machine learning models, specifically focusing on improving forecast accuracy. This paper sought to evaluate an XGBoost ML model's ability to estimate supply chain demand. The researchers used sales data from several departments at 45 Walmart outlets in the United States. Making the data ready for ML by dealing with missing values and converting the categorical features. Data was then splitted into training and testing sets for performance evaluation. An XGBoost produced good results, with an MAE0.0655, an MSE0.1246, and an RMSE0.9838. When comparing the results of XGBoost with other models, such as decision tree regression or linear regression, XGBoost was found to provide better results by a noticeable margin in all three metrics (MAE, MSE, and RMSE). Thus, it can be concluded that XGBoost is an effective approach for enhancing the precision of demand prediction in a supply chain.