Enhanced Classification of Anxiety, Depression, and Stress Levels: A Comparative Analysis of DASS21 Questionnaire Data Augmentation and Classification Algorithms

With mental health concerns increasingly prevalent, particularly among student populations, understanding and addressing these issues are of paramount significance.This study investigates the impact of mental health conditions on individuals within an academic setting, utilizing machine learning models to predict severity based on responses to the DASS-21 questionnaire.This research aims to shed light on the profound impact of mental health on academic communities and to predict and manage mental health conditions effectively.To achieve this, data from 96 participants at American International University -Bangladesh (AIUB) were analyzed using various Machine Learning (ML) algorithms, including random forest, logistic regression, Support Vector Machine (SVM), and gradient boosting.The study found that Logistic Regression demonstrated superior performance over Random Forest, SVM, and Gradient Boosting in terms of F1-score, recall, accuracy and precision.Specifically, logistic regression achieved accuracies of 0.97 for anxiety, 0.975 for depression, and 0.93 for stress.These findings suggest that leveraging machine learning techniques can enhance our understanding of mental health and offer practical avenues for improving well-being and fostering a supportive academic environment.The research holds importance as it has the potential to guide early intervention tactics, offering timely assistance to those requiring support.