Action Based Collaboration Analysis for Group Learning

Mühlenbrock, M.
Pub. date
January 2001
244 of Dissertations in Artificial Intelligence
ISBN print
Artificial Intelligence, Computer & Communication Sciences, Computer Science

With networked computers, new ways of learning and working together with other people are possible. A particular area of computer-supported collaborative learning are shared-workspace systems with structured graphical representations. These systems allow for free user interaction and the joint construction of problem solutions for potentially open-ended tasks. Unlike traditional student modeling as the basis for intelligent support, group modeling in shared workspaces has to take on a process-oriented perspective due to the reduced system control in shared workspaces.
Action-based collaboration analysis is defined as the monitoring of user actions and the abstraction and interpretation of the raw data in the context of the group interaction and the problem representation. Formally based on plan recognition and the situation calculus, an approach (activity recognition) has been developed that incorporates an operational hierarchy for generally modeling activities. The system performs an automatic, online analysis of group interactions, and the results are visualized in different forms to give feedback and stimulating self-reflection.