EXPLAINING HUMAN EMOTIONS USING INTERPRETABLE MACHINE LEARNING FOR BEHAVIORAL AND MENTAL HEALTHCARE

Abstract

The world is encountering an expansion in mental helath problems, as one in each five grown-ups overall as of now experiences mental problems. Physiological, ecological, and natural factors all contribute fundamentally to the advancement of psychological instabilities. Utilizing artificial intelligence techniques enables the development of risk models for assessing an individual's proclivity to develop emotional disorders, consequently improving pre-conclusion screening apparatuses. Notwithstanding, mental and mental medical care are additionally profiting from progresses in AI, for example, PC work for considering, perceiving, and investigating, which can help doctors in distinguishing infections and treating patients suitably. Rather than psychiatrists, robots are being utilized in the cutting edge period to speak with care searchers and suggest treatment choices. The survey features contemporary mechanical progressions to exhibit their evident potential and to give an outline of future advancements. Various functional advantages have likewise been examined following that, which machine innovation brings to mental prosperity care. Humans increasingly use text-based input to share their opinions/emotions whether its about service via online social media or mental health problems. Humans are prone to making erroneous interpretations of emotions, particularly those extracted from the text. The essential target of this study is to foster a feeling acknowledgement and prognostication system that relies on text. The model is based on Ekman's four fundamental emotions: stressed, fear, caring, and loneliness. All models were evaluated against a benchmark dataset of tweeter tweets and reddit comments. A text-based emotion prediction system was successfully developed for the purpose of interpreting and comprehending human emotions that will further help psychiatrists and researchers to tackle mental health disorders.