Performance Analysis of Metaheuristic Methods in the Classification of Different Human Behavioural Disorders

The primary intention of this study is to evaluate the performance of various swarm intelligent algorithms for identifying behavioural disorders. Four separate datasets were investigated and analysed to assess various behavioural problems such as Anxiety, ADHD, and Conduct disorder. Ten different swarm intelligence (SI)-based metaheuristic techniques, viz. slime mould algorithm (SMA), butterfly optimization algorithm (BOA), emperor penguins optimizer (EPO), whale optimization algorithm (WOA), ant lion optimization (ALO), grey wolf optimizer (GWO), firefly algorithm (FF), cuckoo search algorithm (CS), ant colony optimization (ACO), particle swarm optimization (PSO) have been employed to solve feature selection problem for behavioural disorders. The performance of ten distinct methods has been compared based on accuracy, sensitivity, and specificity metrics. The PSO was shown to have the highest accuracy for three of the four datasets.