Mental Health: Prediction and Analysis of Anxiety, Depression, Stress and Happiness Using Machine Learning

Recent years have envisaged mental health as an important parameter in achieving global development goals, as elucidated in United Nations Sustainable Development Goals (UNSDG). Various variables related to negative and positive scales has been identified as Anxiety, Depression, Stress and Happiness. Early analysis and treatment of mental health is of utmost significance to achieve world Happiness Index and with the technical advancement, it has become seemingly easier. This paper emphasizes on classifying and predicting mental health based on four parameters using different machine-learning algorithms. The assessment is done based on K-Nearest neighbors, Random Forest, Logistic Regression, Support Vector Machine and an ensemble model: Gradient Boosting. Comparative results show that the average accuracy rate for Anxiety, Depression, Stress and Happiness of all the algorithms, respectively, was found to be 75.83% (highest) in Support Vector Machine (SVM). This paper also analyzed the outcomes of severity levels of Anxiety, Depression and Stress in an individual and compared it with the Happiness scale to measure the relation between them.