Emotional Health Analysis Using Digital Data and Soft Computing Techniques: A Comprehensive Approach to Understanding Emotional Well-Being

Information gleaned from communication platforms like Twitter may be utilized to assess several facets of mental wellbeing in this research. Due to the rise in consumption of media throughout the COVID-19, the study’s primary objective is to determine if there is a relationship in between online social networking dependence and self-mental wellness. Soft computation approaches and models such as MLP, CNN, and Gaussian Mixture Clustering are thoroughly examined. After collecting 3,890 tweets from attendees in India and the US using the Twitter API for 60 days, we divided those who participated into five categories based on their use regularly: non-problematic, low-risk, moderaterisk, high-risk, and pathological in nature. Key emotions such as happiness, sorrow, anger, and contempt were included in the psychological data collected from the NRC emotion vocabulary. The findings demonstrated that the average amount rose in anger, despair, and other emotions associated with more social media usage. Using the implemented deep learning models, we were able to predict six distinct risk levels associated with mental and behavioural wellness, with usage patterns serving to categorize these levels according to the most prevalent emotions. The results also show that negative emotional or psychological profiles are more likely to be associated with troublesome and heavy online social networking usage, lending credence to the idea that such data might be useful for predicting mental health issues.