A mathematical framework for the automatic creation of artificial neural networks for course authoring systems
J.-F. Arcand · 2002
This article presents a mathematical algorithm that allows the automatic creation of artificial neural networks (ANN) for computerized training environments. The system, HIT, is a computerized training environment composed of five modules. Two modules, Tutorial and Course Author, are directly related to a learner's training. The three remaining modules, Trace Analyzer, ANNs Analyzer and Multidimensional Statistical Analysis, are directly related to the analysis of learner interaction. This article demonstrates the grafting of the mathematical algorithm for creating an ANN in the Tutorial module, using the Course Author module. In this way, the HIT environment allows the author to create course content without needing to know about ANN. As a number of learner types use HIT, the ANN must adapt itself accordingly.