ESAI: An AI-Based Emotional Support System to Assist Mental Health Disorders

Every year more and more people are struggling to handle a wide variety of mental health disorders, such as anxiety and depression. However, as the number of people in need of assistance increases, the number of resources available to them has continued to decrease. The goal of this research is to develop an Emotional Support AI (ESAI) system, an additional resource for those unable to obtain the help and information they need. The ESAI has been trained to classify text based on the Naive Bayes Classification model. The model was trained using 160,000 Reddit posts, which were collected using web scrapping, where users have discussed their experiences with mental health. ESAI provides users with a friendly user-interface from which they can discuss their mental health concerns. The user can choose whether to communicate through typing or through real-time speech recognition. ESAI works by hosting ‘sessions’, in which it will log communications between itself and the user to check for any potential flags that may indicate the user is experiencing symptoms of one or many mental health disorder(s). These sessions can be used by the user for venting or to seek information regarding a variety of mental health disorders. If the probability that the user is experiencing a mental health disorder is higher than a specific threshold, the user is provided with general resources and contacts regarding the specified disorder. The user will also be provided with a mental health evaluation report at the end of each session upon request. Currently, results show that ESAI can classify mental health disorders with seventy-percent accuracy.