A lightweight and unsupervised approach for identifying risk events in news articles

Event identification is important in many different areas of the business world. In the supply chain risk management domain, timely identification of risk events is vital for ensuring the success of supply chain operations. One of the important sources of real-time information from across the world is news sources. However, the analysis of large amounts of daily news cannot be done manually by humans. On the other hand, extracting the related news is very much dependent on the query or the keyword used in the search engine along with the news content. Recent advancements in Artificial Intelligence have opened up opportunities to leverage intelligent techniques for automating this analysis. In our paper, we introduce a lightweight framework that, with only the event's name as input, can autonomously learn all the related phrases associated with that event. It then employs these phrases to search for relevant news and presents the search engine results with a label indicating their relevance. Through this analysis, the framework identifies the event's occurrence in the real world.