Machine Learning Approaches In Demand Forecasting For Supply Chain Management

Machine Learning Approaches

Authors

Vol. 10 No. 03 (2022)
Engineering and Computer Science
March 29, 2022

Downloads

Self-learning systems have wholly revolutionised demand forecasting in the context of SCM due to the high rate of evolution of ML. A problem with most conventional forecasting techniques is that they work for simple exponential progression of demand, but not for most supply chains. The use of the ML techniques including regression models, neural networks, and the compound models will be discussed in improving the forecast accuracy in this paper. This one assesses the effectiveness of these models in responding to SC issues, namely demand volatility and minimizing lead time. This research shows that the use of ML enhances forecasts by a wide margin to reduce costs, increase inventory control and supply chain agility. Suggestions for the incorporation of selected ML tools into previously discussed SCM frameworks are given.

Machine learning system is a revolution to the conventional demand forecasting in a way that it provides supply chain with the relevant tools that can resolve all the existing complexities and uncertainties of demands and supplies. The adoption of machine learning models into SCM practices can provide significant economic returns as well as robustness of operations. Lastly, the paper provides guidelines on how businesses can embrace ML tools, and prospects for research to fill the gaps observed in this practice.