SMHD: a Large-Scale Resource for Exploring Online Language Usage for Multiple Mental Health Conditions

Proceedings of the 27th International Conference on Computational Linguistics, pages 1485–1497 Santa Fe, New Mexico, USA, August 20-26, 2018. 1485 SMHD: A Large-Scale Resource for Exploring Online Language Usage for Multiple Mental Health Conditions Arman Cohan1 * Bart Desmet1,2 * Andrew Yates1,3 * Luca Soldaini1 Sean MacAvaney1 Nazli Goharian1 1IR Lab, Georgetown University, US 2LT3, Ghent University, BE {firstname}@ir.cs.georgetown.edu bart.desmet@ugent.be 3Max Planck Institute for Informatics, DE ayates@mpi-inf.mpg.de Abstract Mental health is a significant and growing public health concern. As language usage can be leveraged to obtain crucial insights into mental health conditions, there is a need for large-scale, labeled, mental health-related datasets of users who have been diagnosed with one or more of such conditions. In this paper, we investigate the creation of high-precision patterns to identify self-reported diagnoses of nine different mental health conditions, and obtain high-quality labeled data without the need for manual labelling. We introduce the SMHD (Self-reported Mental Health Diagnoses) dataset and make it available. SMHD is a novel large dataset of social media posts from users with one or multiple mental health conditions along with matched control users. We examine distinctions in users’ language, as measured by linguistic and psychological variables. We further explore text classification methods to identify individuals with mental conditions through their language. 1 Introduction Mental health is a significant challenge in healthcare. Mental disorders have the potential to tremen- dously affect the quality of life and wellness of individuals in society (Strine et al., 2008; Mowery et al., 2017a). Social media have become an increasingly important source of data related to mental health conditions (Cohan et al., 2017; Mowery et al., 2017b; Coppersmith et al., 2017; Yates et al., 2017), as it is now a prominent platform for individuals to engage in daily discussions, share information, seek advice and simply communicate with peers that have shared interests. In addition to its ubiquity and ease of access, the possibility to disclose mental health matters anonymously or pseudo-anonymously further drives users to online self-disclosure. At the same time, the close connection between language and mental health makes social media an invaluable resource of mental health-related data. Lack of data has been one of the key limitations to understanding and addressing the challenges the domain is facing (Coppersmith et al., 2014a). Data from social media can not only be used to potentially provide clinical help to users in need, but also to broaden our understanding of the various mental health conditions. Social media analysis has already been proven valuable for identifying depression (Coppersmith et al., 2014a; Yates et al., 2017), suicide ideation (Cohan et al., 2017; De Choudhury and Kıcıman, 2017; Kshirsagar et al., 2017; Desmet and Hoste, 2018), and other conditions such as schizophrenia (Mitchell et al., 2015). While social media data is abundantly available, the amount of labeled data for studying mental health conditions is limited. This is due to the high cost of annotation and the difficulty of access to experts. Prior research has investigated self-disclosure as a means of obtaining …

Content truncated to 500 words