Improving Mental Health Diagnosis with Hybrid Ensemble Models: A Data-Driven Approach

Abstract

In today's world, mental health conditions including stress, worry, and depression are more common, particularly among working adults and students. To stop these problems from getting worse, prompt detection and treatment are crucial. This study examines how emotional and behavioural indicators might be used to predict mental health issues using machine learning (ML) algorithms. A mental health dataset was used to train and assess a number of machines learning models, including Logistic Regression, K-Nearest Neighbours, Decision Tree, Random Forest, Gradient Boosting, XGBoost, and a Hybrid ensemble model. Their efficacy was evaluated using performance criteria such F1-score, recall, accuracy, and precision. The Hybrid ensemble approach outperformed conventional algorithms and had the greatest accuracy of 85% among the models that were assessed. The findings show that ensemble approaches have better prediction capacities for mental health detection, especially hybrid strategies. The potential of machine learning to enhance early diagnosis and individualized mental health support systems is highlighted by this study.