People/Web Search Calendar Emergency Info A-Z Index UVA Email   University of Virginia    

Unstructured Information Analysis

Many systems can be accurately characterized using direct observation mediated by electronic sensors. For example, one can collect accurate data for systems ranging from patient health (e.g., blood pressure monitoring) to transportation infrastructure and social networks. However, many human-oriented systems also contain a tremendous amount of information stored as unstructured text. In all such systems, the complexity and ambiguity of natural (i.e., human) language limit our ability to perform automated observation and analysis. Our research removes these limitations by automatically identifying structured information within unstructured text. Research in this area has included the following:

  1. Predicting future crime based on analyses of social media content
  2. Extracting medical knowledge from clinical practice guidelines for automated decision support
  3. Detecting terrorist recruitment in online discussion forums

Primary Faculty:

Donald E. Brown

Matthew S. Gerber