Abstract
Online mental health forums provide users with an anonymous support platform that is facilitated by moderators responsible for finding and addressing critical posts, especially those related to self-harm. Given the seriousness of these posts, it is important that the moderators are able to locate these critical posts quickly in order to respond with timely support. We approached the task of automatically triaging forum posts as a multiclass classification problem. Our model uses a supervised classifier with various features including lexical, psycholinguistic, and topic modeling features. On a dataset of mental forum posts from ReachOut.com, our approach identified critical cases with a F-score of over 80%, showing the effectiveness of the model. Among 16 participating teams and 60 total runs, our best run achieved macro-average F1-score of 41% for the critical categories (The best score among all the runs was 42%).
Cite
CITATION STYLE
Cohan, A., Young, S., & Goharian, N. (2016). Triaging mental health forum posts. In Proceedings of the 3rd Workshop on Computational Linguistics and Clinical Psychology: From Linguistic Signal to Clinical Reality, CLPsych 2016 at the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL-HLT 2016 (pp. 143–147). Association for Computational Linguistics (ACL). https://doi.org/10.18653/v1/w16-0316
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
