Loading...
Developing computational models for spiritual care in online support communities
Link, Noah
Link, Noah
Citations
Altmetric:
Advisor
Editor
Date
Date Issued
2025-04
Date Submitted
Keywords
Collections
Research Projects
Organizational Units
Journal Issue
Embargo Expires
Abstract
As mental health has become more understood as an integral part of wellness, spiritual health–or one’s sense of meaning and purpose–is still largely neglected. Spiritual care providers, or chaplains, have assiduously provided this form of care through providing immediate human connection that can aid patients to grapple with questions of meaning, purpose, and mortality. Although human-computer interaction research has explored the ways how online communities buttress individuals through difficult times, there remains a gap in the understanding of chaplains within these communities. Thus, the aim of this research is to classify users based on their original posts into categories regarding their role through machine learning models like a LSTM with GloVe embeddings and fine-tuning a BERT model. Of the 386 users classified with the final model with 84% cross-validation accuracy, we find that chaplains (48 users) and medical professionals (79 users) are extremely difficult to classify due to conflation with users who may use spiritual or medical concepts in general terms, which led to many false positives. Moreover, we note the common occurrence of people exhibiting symptoms (273 users), receiving care (42 users), and expressing care (127 users) throughout these communities which all were classified with high accuracy. Finally, we lay the groundwork for future inquiries into the behaviors of different roles within these communities through computational methods.
Associated Publications
Rights
Copyright of the original work is retained by the author.