Rapid Evaluation and Analysis of Communication in Teams (REACT)

In a Net Centric Battlespace, information is constantly shared along both formal and informal channels. Informal means of communication such as chat rooms are often employed when standard channels of communication do not fit the emergent work practice of a group. Despite its inefficiency, this unstructured communication is very natural and convenient for users. As such, it is a rich and plentiful data source for understanding group interaction. However, existing methods for analyzing such discourse are either low-fidelity (simple keyword matching) or extremely man-power intensive (manual transcription and tagging). With advances in speech recognition, part-of-speech tagging, and social networking theory, it is possible to improve on these discourse analysis techniques and use them to model an interaction. One primary output of this model is the performance of individual team members, and of the team as a whole. Such measures can be used to highlight areas of reduced performance or possible breakdown within a large team. We propose to develop a system for Rapid Evaluation and Analysis of Communication in Teams (REACT), combining traditional techniques from discourse analysis with advances in NLP and communication analysis. REACT rapidly analyzes streams of communication to highlight potential areas of coordination difficulty.