Mental Health Diagnostics Using Machine Learning: A Clustering and Predictive Modeling Approach

The application of machine learning has rapidly advanced as a transformative approach in healthcare, offering significant potential in the analysis of huge amounts of acquired data on Mental health across the world enhancing the treatment of mental health issues. Machine learning algorithms are being utilized to analyse patterns and trends in mental health datasets. This study investigates various machine learning methodologies and descriptive analytics techniques to detect patterns and trends in mental health-related factors, focusing on predicting the likelihood of individuals seeking mental health treatment. We utilized clustering, association rule mining, feature selection, and machine learning classifiers such as Random Forest, Decision Tree, XGBoost, and Logistic Regression. By applying these techniques to a dataset of mental health factors, we identified key predictors influencing mental health outcomes and developed models to enhance predictive accuracy. Our analysis highlighted the effectiveness of machine learning algorithms in extracting meaningful patterns from the dataset, with Random Forest and XGBoost emerging as particularly strong performers with accuracies of 81% and 84%. Other applied algorithms that are Decision Tree, Gradient Boosting and logistic regression have the accuracy of 81%, 77%, and 73% respectively.