Flexible User-Guidance in Multimedia CBT-Applications Using Artificial Neural Networks

The paper decribes a system architecture for CBT-applications that includes artificial neural networks (ANNs) for dynamically creating user-specific paths through a database of multimedia objects. The organisation of these objects is based on principles of hypersystems. Additionally a semantic data model is involved to describe the objects’ characteristics. The ANNs represent experiences and decisions of former users, based on the object’s characteristics. During a CBT-session, the ANNs are used by a run-time controller in order to retrieve objects, which are appropriate candidates for continuing the lesson, from the semantic data model.