Mental Health Awareness Using Machine Learning

Mental disorders are a serious issue that affects the well-being of millions of people across the world. Mental health is common however it is often left undiagnosed and untreated which is largely due to social stigma, which, even in today's world, is not commonly understood or talked about. In addition, we have a lack of economic resources to help fund research and give out the proper treatments that are needed. This research proposes the implementation of a machine learning system to help increase the common knowledge about mxental disorders and help with self-diagnosis. The system contains three primary parts: (1) predicting stress from behavior and physiology data using the XGBoost algorithm and training it on the basis of those data; (2) detecting emotions through a Convolutional Neural Network (CNN) that classifies facial expressions; and (3) a digital mood diary that the user can fill out to record and monitor their long-term emotional patterns. By integrating these three modules, the user is provided with information about their mental state that can promote a greater level of self-awareness and elicit the kind of early intervention that can make a significant difference. The findings of this research support the creation of accessible datainformed mental health resources that can help increase user engagement and reduce the barriers to early detection.