Unapparent information revelation: Text mining for counterterrorism

Unapparent information revelation (UIR) is a special case of text mining that focuses on detecting possible links between concepts across multiple text documents by generating an evidence trail explaining the connection. A traditional search involving, for example, two or more person names will attempt to find documents mentioning both these individuals. This research focuses on a different interpretation of such a query: what is the best evidence trail across documents that explains a connection between these individuals? For example, all may be good golfers. A generalization of this task involves query terms representing general concepts (e.g. indictment, foreign policy). Previous approaches to this problem have focused on graph mining involving hyperlinked documents, and link analysis exploiting named entities. A new robust framework is presented, based on (i) generating concept chain graphs, a hybrid content representation, (ii) performing graph matching to select candidate subgraphs, and (iii) subsequently using graphical models to validate hypotheses using ranked evidence trails. We adapt the DUC data set for cross-document summarization to evaluate evidence trails generated by this approach