A study on mental health discussion through Reddit

• Main Authors: Nur Shazwani, Kamarudin, Beigi, Ghazaleh, Liu, Huan
• Format: Conference or Workshop Item
• Language: English
• Published: IEEE 2021
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://umpir.ump.edu.my/id/eprint/43576/1/A_Study_on_Mental_Health_Discussion_through_Reddit.pdf; http://umpir.ump.edu.my/id/eprint/43576/; https://doi.org/10.1109/ICSECS52883.2021.00122

The massive growth of social media for the past few decades has given a new era to the open web. People are using social media to openly share opinions and also discussed sensitive subjects such as mental health. This paper study the online community on the Reddit discussion forum. We focused on studying linguistic behavior in the online mental health community. We report linguistic analysis in order to understand the similarity and differences of each mental health community. We first study the sentiment per each Reddit community and followed by the topic modeling. By using the well-known Latent Dirichlet Allocation (LDA) method, we extract the most discussed topic for each community. By utilizing all related posts from each subreddits in each community, we establish the following exciting insights: (a) We observe that the headline of the post does not indicate the whole content of the discussion post, (b) we found that sentiment across the mental health community was high on positive compared to negative, and (c) topic distribution for each community varies but there exist similarities among them.