A hybrid particle swarm optimization-support vector machine for anxiety prediction among undergraduate students.

Abstract

Globally, the prevalence of anxiety disorders is increasing, and this illness has a significant impact on people's mental health, cognitive functions, behavior, and social interactions, among other aspects of their lives. Anxiety disorders throw off the delicate balance of mental health. Exploring how machine learning can be harnessed to combat the decline in life expectancy by addressing anxiety health issues is a vital issue. Novel research demonstrates the revolutionary possibilities of unconventional methods. These methods provide a deeper understanding of mental health issues by analyzing the complexity of these disorders and the underlying causes. Consequently, this research aims to explore the role of Artificial Intelligence (AI) in the assessment of anxiety levels among Nigerians. By leveraging AI technology, we can develop a more accessible and proactive approach to mental health monitoring and support. Metaheuristic algorithm: Particle Swarm Optimization algorithm is being explored as optimization techniques to increase the performance accuracy of these AI methods and overcome some challenges of traditional AI models. Our findings indicate that the Oversampled, PSO-tuned Support Vector Classifier outperformed the other models with 96% Accuracy, 96% Precision, 99.8% Recall, and F1 score of 83.3%. Finally, this study presents a Hybrid Particle Swarm Optimization-Support Vector Machine for Anxiety Prediction among Undergraduate Students and indicates that oversampling increases minority class instances to balance class distribution and overcome imbalanced datasets