Text analytics for supply chain risk management

Abstract

Supply chain risk management has always been a crucial field in both academia and industry. The globalization of supply chains changes the way companies operate and coordinate with their supply chain stakeholders. The increasing involvement of the companies worldwide results in win-win benefits for different tiers of supply chains, but meanwhile, the growing complexity of it makes it more challenging to manage the risks. This dissertation includes multiple studies about supply chain risk management using open-access textual data collected online. A literature review of the web media applications in stock markets is conducted to exhibit how different text analytics methods can be used in stock market analyses. A global supply chain risk management framework is proposed using correlation analysis and n-gram representation to develop a risk categorization. Sentiment analysis of online news articles is also conducted for capturing the risk pattern variation. One of the studies also incorporates factor analysis and analytic network process to deal with the interdependency of risk factors and develop a document-count-based scoring method for the resource limits of a company. Finally, the relationship of the document-count-based index, Global 500 business ranking, and sentiment polarity is explored to validate the effectiveness and potential of the scoring method.