Comparative Analysis of Different Meta Heuristic Optimization Algorithms in Depression Detection System

Depression is one of the most common mental health conditions that affect people of all ages around the world today. Depression not only poses diagnostic challenges for clinicians and medical professionals, but it is also socially unacceptable and connected with prominent social stigmas. Researchers have been investigating how to use computers and automated systems to detect depression symptoms in people based on their speech and behaviours. This paper performs a comparative analysis of different Metaheuristic optimization algorithms in depression detection systems. This paper proposes a parallel gated convolutional capsule transformer (PGCCtrans) model for detecting depression through speech. Evaluation results show that PGCC-trans model + BHOA, PGCC-trans model + CDMOA, PGCC-trans model + DBOA, PGCC-trans model + GJOA and PGCC-trans model + HOA attained accuracies of 99.56%, 99.529%, 99.61%, 99.52%, and $\mathbf{9 9. 6 0 \%}$ respectively. The findings demonstrate the effectiveness of the PGCC-trans model and metaheuristic optimization algorithms for speech-based depression detection.