CAMS: An Annotated Corpus for Causal Analysis of Mental Health Issues in Social Media Posts

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 6387–6396 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 6387 CAMS: An Annotated Corpus for Causal Analysis of Mental Health Issues in Social Media Posts Muskan Garg1,7, Chandni Saxena2, Veena Krishnan3, Ruchi Joshi4, Sriparna Saha5, Vijay Mago6, Bonnie J Dorr7 1Thapar Institute of Engineering & Technology, India, 2The Chinese University of Hong Kong, Hong Kong SAR, 3University of Petroleum And Energy Studies, India 4Amity University Rajasthan, India, 5Indian Institute of Technology, Patna, 6Lakehead University, Canada, 7University of Florida, USA. {muskangarg, bonniejdorr}@ufl.edu, chandnisaxena@cuhk.edu.hk, sriparna@iitp.ac.in, vkrishnan@ddn.upes.in, rjoshi@jpr.amity.edu, vmago@lakeheadu.ca Abstract Research community has witnessed substantial growth in the detection of mental health issues and their associated reasons from analysis of social media. We introduce a new dataset for Causal Analysis of Mental health issues in Social media posts (CAMS). Our contributions for causal analysis are two-fold: causal interpretation and causal categorization. We introduce an annotation schema for this task of causal analysis. We demonstrate the efficacy of our schema on two different datasets: (i) crawling and annotating 3155 Reddit posts and (ii) re-annotating the publicly available SDCNL dataset of 1896 instances for interpretable causal analysis. We further combine these into the CAMS dataset and make this resource publicly available along with associated source code: https://github.com/drmuskangarg/CAMS. We present experimental results of models learned from CAMS dataset and demonstrate that a classic Logistic Regression model outperforms the next best (CNN-LSTM) model by 4.9% accuracy. Keywords: clinical depression, clinical psychology, intent classification, suicidal tendency 1. Introduction With substantial growth in digitization of psychologi- cal phenomena, automated Natural Language Process- ing (NLP) has been applied by academic researchers and mental health practitioners to detect, classify or predict mental illness on social media. However, there is a critical need for identifying underlying causes of mental illness in the face of dire outcomes. For ex- ample, a person commits suicide every 11.1 minutes in the US1 and 23% of deaths in the world are associated with mental disorders according to the World Health Organization. The pandemic lockdown has heightened the mental health crisis in UK (Pierce et al., 2020) and US (McGinty et al., 2020). In this context, people with mental disorders who decide to visit mental health practitioners for social well-being may face difficulty due to social stigma or unavailability of mental health practitioners, leaving those most in need to be neglected by the community. As a result, sufferers of mental health conditions are unable to take necessary steps for their treatment. Un- fortunately, 80% of those with mental health conditions do not undergo clinical treatment and about 60% of those who take their own lives previously denied hav- ing any suicidal thoughts (Sawhney et al., 2021). Ac- cordingly, social media platforms (e.g., Reddit, Twit- ter) are important resources for investigating the men- tal health of users based on their writings. 1https://suicidology.org/wp- content/uploads/2021/01/2019 datapgsv2b.pdf Level 0: Classification/ Longitudinal study Level 1: Interpretable (direct) Causal Analysis Level 2: Identifying latent cause …

Content truncated to 500 words