Improving the Resilience of Supply Chains in a Post-COVID-19 Era

Abstract

The COVID-19 pandemic has highlighted the critical need for supply chain resilience in the face of unforeseen disruptions. This research investigates the application of machine learning (ML) algorithms to enhance supply chain resilience during the COVID-19 crisis. The authors evaluated several ML algorithms, including decision trees, random forests, naïve bayes, and LSTM. They explored using the SPIN COVID-19 RMRIO dataset to develop a proactive and data-driven approach to mitigate disruptions and improve supply chain performance. The ML model worked with and without feature selection. With chi-square feature selection, the long short-term memory (LSTM) performed well and achieved the highest accuracy, 96.74%, with an F1 score of 91.01%. Without feature selection, random forest outperformed, which provided an accuracy of 96.21% with an F1 score of 81.25%.