Depression Detection Using Sentiment Analysis Of Social Media Posts

Abstract

Depression has become a major cause of concern worldwide. According to the World Health Organization (WHO), more than 264 million people of all ages suffer from depression worldwide. Almost 75% of these remain untreated, with almost 1 million people taking their lives each year. This makes depression one of the leading causes of suicide esp. amongst adolescents.  The WHO reports that anxiety disorders are the most common mental disorders worldwide with specific phobia, major depres- sive disorder and social phobia being the most common anxiety disorders. Social media platforms are becoming an integral part of people’s life. They reflect the user’s personal life. People like to share happiness, joy, and sadness on social  media.  These  platforms are used by researchers to identify the causes of depression and detect it. Detecting early depression can be a huge step to address this mental illness. Our model addresses this issue by deploying a hybrid machine learning model using XGBoost and Naive Bayes, which would help in detecting individuals with symptoms of clinical depres- sion. The model will be developed for twitter data and would flag tweets which are found to be depressive. The project is expected to give an accuracy of 90% and a very good F1-Score. Thus,   this project would help us employ emotional AI in twitter which would in turn lead to lower suicide rates and improved mental health.