Modeling the Impact of the Covid-19 Risks on Global Supply Chains using Text Mining Methods: A Case Study of the Food Supply Chain

The first step in analyzing and evaluating supply chain risks is to identify these risks. Conventional analysis methods are based on manual filters or data-driven automated methods. Manual filtering methods suffer from validation problems due to sampling limitations, and automated data analysis methods, on the other hand, perform poorly in analyzing risk data that is complex and ambiguous. To fill the research gap, in this study, an interactive framework between the analyst and the machine is presented to analyze a large volume of risk data in the field of food supply chain using topic modeling techniques, word embedding, term correlation analysis, and knowledge map. The purpose of the supply chain risk monitoring system is to assist managers in food companies to monitor and identify crisis risks and provide a decision support information system to create a sustainable food supply chain. The results of the topic analysis of metadata showed a knowledge map in five areas of "harvest", "agriculture", "food retail", "food services", "distribution" and "consumption", which was approved by the expert panel. The results show that the risk analysis model is useful in extracting knowledge units related to the field of food supply chain crisis management.