ABSTRACT
The COVID-19 pandemic continues to negatively impact people's mental health worldwide. Due to the rise in unemployment, loss of income, and lack of social interaction, people are now more likely to feel lonely, go on fewer outings, and dread the unexpected nature of viral transmission. Meanwhile, Public Health authorities are interested in monitoring people's mental and emotional well-being. In this paper, natural language processing is used to analyze human sentiments concerning the COVID-19 pandemic that has been dangerously affecting individuals' mental and physical well-being for more than two years now. Even though several waves of COVID-19 have passed, of which the first and third waves i.e., the initial pandemic period from 20th March 2020 and the rise of the Delta variant from January 2020 had the most impact on the mental health of individuals, this is further evident by the results of this paper. This research focuses on how severely this virus has affected people's mental health and emotions. After processing the data i.e., cleaning, formatting, and removing irregularities from the data, feature engineering models are applied to acquire the results. The results through VADER (valence-aware dictionary and sentiment reasoning) indicate an increase in overall negative sentiments between two mentioned periods. Additionally, the NRC-EIL (National Research Council of Canada - Emotion Intensity Lexicon) analysis showed that 'fear' and 'sadness' occurred during those times. © 2022 IEEE.