Multi-Model Machine Learning Approach for Supporting Sri Lankan Veteran Mental Health
I. S. C. Dharmapriya, S. G. B. V. R. Pathirana, H. G. Hettiarachchi, Dilan Warnakulasooriya, Samantha Thelijjagoda, Jenny Krishara, Nishantha Giguruwa · 2024
Sri Lankan veterans face unique mental health challenges that demand culturally sensitive care. This research introduces an innovative AI-based mental health support system to meet these needs. Our system integrates four components for comprehensive, personalized support. First, a deep learning model trained on a synthesized dataset (Synthea) diagnoses mental health disorders. This diagnosis, combined with responses to standardized questionnaires (e.g., PHQ, DSM-V), enables the second component – a machine learning model – to recommend tailored Cognitive Behavioral Therapy (CBT) interventions. Additionally, real-time sentiment analysis of the user’s voice and a text-based emotion detection model provides insights into current emotional states for refined recommendations. These AI models work seamlessly within a user-friendly mobile application that continuously adapts its recommendations based on user input. This system holds the potential to transform mental health support for Sri Lankan veterans, providing culturally responsive care that continuously evolves alongside their needs.