An Overview of the Application of Sentiment Analysis in Mental Well-being

Abstract

Sentiment analysis is a field of study in natural language processing, usually implemented by machine learning and deep learning methods. Its purpose is to identify the information from pieces of text and categorize them based on their subjective opinion, such as positive, negative, and/or neutral. It has been used in a variety of fields, usually fields related to business, such as customer service and marketing, to analyze the opinions of customers to help businesses adjust their products to accommodate the customers needs. This work analyzes the use of sentiment analysis in mental health. Starting in 2020, the emergence of COVID-19 has disrupted the lives of billions around the world. The isolation brought by the pandemic has forced the public to spend a much greater amount of time on social media and very little time socializing. This has caused a significant increase in the numbers of mental diseases such as anxiety, depression, obsessive-compulsive disorder (OCD), schizophrenia, and bipolar disorder. This paper provides an overview of the recent research of machine learning and deep learning methods on sentiment analysis in the field of mental health since 2020, when the COVID-19 was confirmed and obligatory and non-obligatory quarantines were enforced within different countries. Then, this paper examines the potential applications of sentiment analysis in addressing the mental illness of the public, in aspects of diagnosing feedback from patients, social media monitoring, and mental health apps. Lastly, several challenges that sentiment analysis still faces today will be presented and briefly analyzed. More research on sentiment analysis needs to be done in order to achieve a higher level of accuracy when applying sentiment analysis models to psychiatric uses, but the benefits of a developed sentiment analysis model is highly substantial in mental health.