Abstract:
Attention Deficit Hyperactivity Disorder
(ADHD) is one of the common psychiatric disorder in childhood,
which can continue to adulthood. The ADHD diagnosed
population has been increasing, causing a negative impact on
their families and society. This paper addresses the effective
identification of ADHD in early stages. We have used a rulebased
approach to analyse the accuracies of decision tree
classifiers in identifying ADHD subjects. The dataset consists of
eye movements and eye positions of different gaze event types.
The feature extraction process considers fixations, saccades,
gaze positions, and pupil diameters. The decision tree-based
algorithms have shown a maximum accuracy of 84% and
classification rule algorithms have shown an accuracy of 82%
using eye movement measurements. Thus, both algorithms have
shown high accuracy with the rule-based component.
