Story extraction from the Web: A case study in security informatics

Open source intelligence is becoming more and more important in security-related applications. Effective extraction of valuable information from these intelligence sources so as to understand the content of intelligence is one of the central issues in security domain. To address this key problem, this paper studies how to extract domain events and generate story representation of the events. We implement a story extraction platform (SEP) which applies pattern matching to extract events from Web news and organizes events based on theme using narrative structure. We also design extraction rules, use domain-specific features and employ ontology to facilitate story extraction in SEP. The experimental results show the effectiveness of our system in security informatics.