Model-based optimisation for enhanced training of individuals based on abilities, learning styles and preferences
Date
2013-01-29
Authors
Govender, Viren
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Computer based training of individuals is becoming more common. Computer based
systems increasingly are filled with devices and appliances that enhance the user’s interaction
with the computer. These new devices and appliances present new modalities
of interaction with the user. This opens new possibilities for computer based training.
However, not much is known about mapping these modalities to the user for enhanced
learning. This thesis presents an artificial learning model for on-line training of individuals.
The model supplied is a multi-modal system in that it links multiple input and
output modalities to a user profile. The model contains a non-linear mapping between
the user profile and the modalities. The non-linear mapping has been achieved through
the use of an Artificial Neural Network. The learning model has been extended to include
time dependencies of the suggested modalities via a feedback mechanism within the Artificial
Neural Network. The presented results indicate the complexity in choosing the most
appropriate mapping for an individual. Results are presented showing the robustness of
the learning model. By taking cognisance of the user profile and context (e.g. the user is
bored or tired) appropriate modalities are suggested which facilitate learning.