APPLICATION OF MACHINE LEARNING IN SUPPLY CHAIN AND MANAGEMENT

Abstract

Day by day supply chain is becoming more and more competitive. It is now necessary for organisations to accurately predict their customers' behaviour to deal with unexpected surges and lull in the demands of product. With new methods of data collection coming up the volume of data generated is large in number as well as in variety. So, it becomes difficult for traditional methods to make forecasts, for the same reason, the idea of using more advanced Machine Learning techniques is being explored. For the dissertation a comparative analysis is done to evaluate the performance of four model namely – Linear Regression, Neural Network, Random Forest and Decision Tree. All these four models are applied to a common case study of forecasting demand for a meal delivery company. Different aspects of application of these models are discussed in the methodology section. Three performance metrics are used to evaluate the performance and to find out which model gives best result for this particular dataset.