An AI-Based Framework for Interpretable Mental Health Literacy Segmentation and Decision Support

Mental Health Literacy (MHL) is a multidimensional aspect that addresses barriers to seeking professional help. Strong MHL enables individuals to identify mental health challenges and seek professional support if required. Although its importance is widely recognized, there is a significant gap in the design of targeted interventions considering modifiable factors. This study offered a decision support framework that assessed multiple dimensions of MHL and segmented individuals within academic communities based on modifiable predictors that can be amenable to change through educational interventions. The user study was conducted using a cross-sectional survey of 385 participants, representing diverse academic roles. Five latent profiles were identified, each reflecting distinct configurations of help-seeking attitude, knowledge, and beliefs related to mental illness using Latent Profile Analysis. The study compared decision tree classifiers (C5.0, CART, and Conditional Inference Trees) to generate decision rules tailored to each profile, ensuring interpretability and practical utility. Interactions between MHL profiles and demographic variables are visualized through an interactive POWER BI dashboard. This framework provides actionable insights to policymakers and health professionals for targeted resource allocation, precise intervention design, and longitudinal tracking of profile evaluation, thereby enhancing institutional mental healthcare.