Human Mental Health Prediction using Machine Learning Models

Susmitha Karanam, Masters of Research Methods in Psychology, University of Strathclyde,Glasgow,UK
Shaik Jumlesha, Professor, Annamacharya Institute of Technology,Tirupathi,India
susmitha.karanam.2025@uni.strath.ac.uk1 ahmedsadhiq@gmail.com2

ABSTRACT:Early detection of mental health issues enables specialists to provide more effective treatments, significantly enhancing patients' quality of life. Mental health encompasses psychological, emotional, and social well-being, influencing how individuals think, feel, and behave. It is vital throughout all life stages, from childhood and adolescence to adulthood. This study focused on identifying and evaluating the accuracy of five machine learning techniques in detecting mental health issues. The techniques analyzed were Logistic Regression, KNearest Neighbours (K-NN) Classifier, Decision Tree Classifier, Random Forest, and Stacking. We assessed their performance based on several accuracy criteria to determine the most effective method. After comparing these techniques, we found that the Stacking technique, which combines multiple models to improve prediction accuracy, achieved the highest accuracy rate at 82.75%. This indicates that Stacking is a promising approach for accurately identifying mental health issues, potentially leading to better early intervention and treatment outcomes. Keywords: Mental health, Early Detection, Machine learning Techniques, Accuracy Evaluation, Health Awareness, Mental Wellbeing.