A neural approach for modeling the inference of awareness in computer-supported collaboration

Abstract

Individuals interacting in a computer supported collaborative learning (CSCL) environment produce a variety of information elements during their participation; these information elements usually have a complex structure and semantics, which make it rather difficult to find out the behavioral attitudes and profiles of the users involved. This work provides a model that can be used to discover awareness information lying underneath multi-user interaction. This information is initially captured in log files and then is represented in a specific form in events-databases. By using data mining techniques, it is possible to infer both the users' behavioral profiles and the relationships that occur in a CSCL environment. In this work we combine different data mining strategies and a neural-based approach in order to construct a multi-layer model that provides a mechanism for inferring different types of awareness information from group activity and presenting it to the interested parties. © Springer-Verlag Berlin Heidelberg 2006.