An evolutionary framework on ADHD diagnosis based on graph theory and ant colony optimisation

Abstract

Developing countries facing unavoidable issues for the parents living with children due to attention deficit hyperactivity disorder (ADHD). This neuropsychiatric disorder has effects on the children in terms of inattentive, impulsive, and hyperactivity. Graph theory provides useful description measures as predicted vectors for the classification process and this research work provides an automated diagnosis model for predicting the ADHD features based on the neural network classifier to differentiate ADHD patients and their healthy controls from a combined environment includes normal persons and affected patients. Ant colony optimisation model is used to get converged results for the classifier results in terms of both phenotypic data and imaging data. ADHD-200 dataset is used for analysis in the proposed model. The experimental result yields an accuracy of 86% on two class diagnosis better than phenotypic approaches.