A Machine-Learning Approach to Classify Depressive Posts from Reddit Social Media

Depression leading to suicide has become a major concern in our society. It's high time to investigate the reason behind it. As numerous people are expressing their concerns regarding depression on Reddit, such post can be taken into consideration in order to find the reason behind it. There has been a lot of research work in this field but most of it is done using traditional machine learning approach with NLP techniques. Moreover, recently there's has not been any significant research work regarding depression. In this paper, we found the indicators that could indicate how relevant Reddit users feel about depression using machine learning, ensemble learning, and deep learning techniques. Moreover, we pinpointed the terms that appear more frequently in narratives of depression. To strengthen the research work, we took the dataset over the period from 2017 to 2021. Furthermore, we demonstrated the effectiveness of LSTM & BERT on recent dataset.