Title: Mining critical events from news using NewsReader
The NewsReader project created software to automatically detect events in news texts to generate Event-Centric-Knowledge-Graphs or so-called ECKGs. ECKGs are Semantic Web representations of what happened, who was involved, where and when so that computers can reason over the data. Entities are linked to structured databases such as DBPedia (based on Wikipedia) and events are linked to the Circumstantial Event Ontology (CEO). In this presentation, I will explain the NewsReader method, which is based on Identification, Deduplication, Aggregation and Perspectivation (IDAP), and show how the CEO can be used to detect causal series of events. Causal event series explain us why things happened and what things are likely to happen given a described situation in the news. Eventually, such technology can be used for risk analysis in for example supply chain dependencies of industries.