Kavuluru, Ramakanth, et al. "Classification of helpful comments on online suicide watch forums." Proceedings of the 7th ACM international conference on bioinformatics, computational biology, and health informatics. 2016.

This study explores the automatic identification of helpful comments in Reddit's Suicide Watch (SW) subreddit to assist human moderators and improve suicide prevention (SP) efforts. Using a dataset of 3000 comments, supervised learning experiments with features like n-grams and psychometric scores achieved high accuracy in identifying helpful comments. The findings suggest that machine learning can complement human moderation and highlight the need for further research in assessing comment helpfulness for mental health posts.