This study proposes an interpretable artificial intelligence framework for mental health literacy (MHL) stratification using Latent Profile Analysis (LPA) and Decision Tree classification models. The framework is validated on the publicly available OSMI Mental Health in Tech Survey dataset comprising 1,259 anonymized responses from professionals across organizational settings. The proposed approach identifies five distinct mental health literacy profiles based on behavioral and psychological indicators, including help-seeking attitudes, perceived workplace support, stigma toward mental health disclosure, and psychological work interference. The C5.0 decision tree model achieved the highest classification performance with 84% accuracy, demonstrating the effectiveness of interpretable machine learning for real-world mental health analytics. The framework provides actionable insights for policymakers and organizational decision-makers to design targeted interventions, allocate resources efficiently, and support evidence-based mental health programs. This work contributes to the advancement of explainable AI in public health and organizational mental health management.
Lakshya Gupta (Mon,) studied this question.