APPLICATION OF MACHINE LEARNING IN SUPPLY CHAIN MANAGEMENT – LEVERAGING SUPERVISED MACHINE LEARNING FOR EFFICIENT SUPPLIER SELECTION

Abstract

The aim of this paper is to discuss the inefficiencies in supplier selection which occur in the shipping and logistics industries. These inefficiencies interrupt the regular flow of goods in the global economy and obstruct optimal performance. Traditional methods that rely on human review and subjective standards sometime leads to poor selection and missed opportunities. The purpose of this work is to introduce new methods that use supervised machine learning(SML) algorithms to increase the accuracy and efficiency of vendor selection process. With the aim of enhancing selection accuracy through the use of supervised machine learning algorithms in our research, "Application of Machine Learning in Supply Chain Management – Leveraging Supervised Machine Learning for efficient Supplier Selection" focuses on the supplier selection process based on available features and instances. This research uses a two-phase methodology. Initially, we examine historical information from previous procurement activities, encompassing vendor performance metrics. Second, we create a predictive model for vendor selection using a variety of machine learning algorithms, including Random Forests, Decision Tree, and KNearest Neighbor models. These algorithms were trained and compared to find the model with best accuracy in forecasting vendors for upcoming contracts. While comparing the different Supervised Machine Learning (SML) models and SML with sparrow search algorithm as feature selection technique, it has been found that, the accuracy of vendor selection was significantly higher for the Random Forest which is 92.69% followed by the decision tree 86.4% KNearest Neighbor(KNN) 85.4%, Adaboost 84.21%. The random forest with the SSA algorithm's accuracy was the highest. The study also emphasizes how important it is to train machine learning models using large data sets. The shipping logistics industry can greatly optimize its resources by integrating machine learning algorithms with data analytics. This study demonstrate how the application of supervised machine learning algorithms can fundamentally alter the process of vi selecting suppliers for logistics and transportation projects. Predictive modelling and historic data can help organisation make well informed decisions, reduce disruptions, and enhance the overall effectiveness of their supply chain operations. The study findings indicate that more research is needed into merging SML algorithms optimization techniques to consistently improve supplier selection’s accuracy. This would boost operational efficacy and industrial competitiveness.