DEMAND FORECASTING USING MACHINE LEARNING

This study examines the importance and approaches of demand forecasting in the context of digitalization of business processes. This paper examines the differences between traditional statistical methods (Moving Average) and machine learning methods (Random Forest). Through experimentation using the Python programming language and a real-world retail dataset, it was found that machine learning techniques achieved significantly higher accuracy (MAE) than traditional methods. The paper concludes by highlighting the importance of new technologies in preventing overstocking and improving cost efficiency in large retail and logistics companies.