Suicidal Comment Tree Dataset: Enhancing Risk Assessment and Prediction Through Contextual Analysis
Jun Li, Qun Zhao
TL;DR
The paper tackles the limitation of post-only suicide-risk detection by introducing a longitudinal, comment-tree–based approach. It builds a Reddit-based Suicidal Comment Tree Dataset with two historical post sequences per case and refines a C-SSRS-derived four-label scheme to annotate risk levels, validating reliability with high inter-annotator agreement. Through targeted experiments using Chain-of-Thought prompting across multiple LLMs, it shows that including comment-tree context improves discrimination and prediction of current risk, with model- and data-dependent gains. The work offers a publicly available dataset and a practical annotation framework, highlighting the potential for earlier, more accurate intervention while acknowledging platform-specific limitations and ethical considerations.
Abstract
Suicide remains a critical global public health issue. While previous studies have provided valuable insights into detecting suicidal expressions in individual social media posts, limited attention has been paid to the analysis of longitudinal, sequential comment trees for predicting a user's evolving suicidal risk. Users, however, often reveal their intentions through historical posts and interactive comments over time. This study addresses this gap by investigating how the information in comment trees affects both the discrimination and prediction of users' suicidal risk levels. We constructed a high-quality annotated dataset, sourced from Reddit, which incorporates users' posting history and comments, using a refined four-label annotation framework based on the Columbia Suicide Severity Rating Scale (C-SSRS). Statistical analysis of the dataset, along with experimental results from Large Language Models (LLMs) experiments, demonstrates that incorporating comment trees data significantly enhances the discrimination and prediction of user suicidal risk levels. This research offers a novel insight to enhancing the detection accuracy of at-risk individuals, thereby providing a valuable foundation for early suicide intervention strategies.