ABSTRACT
Given the enormous volume of text data currently available through online sources, intelligence analysts require tools to identify key entities and understand their relationships.The entity-relationship topic (ERT) model provides a mechanism for extracting novel entity characteristics and relationships from text. We report on the application of the ERT model to an Egyptian news dataset and also describe visualizations that can intuitively display the outputs of the model. Starting with known persons of interest, we can identify their key characteristics and understand the relationships that they share with other entities. Alternatively, starting with a topic of interest, the model outputs show the network of players associated with the topic, and the strength of that association. We thus demonstrate that the ERT model can provide an analyst with a quick understanding of relationships within an area of operations.
