Holistix: A Dataset for Holistic Wellness Dimensions Analysis in Mental Health Narratives

2025-07-13

AI Summary

Prior work lacks fine-grained, multidimensional frameworks for assessing mental health from social media text grounded in comprehensive theoretical models. Method: We propose the first multidimensional fine-grained classification framework for mental health grounded in the six-dimensional wellness model (physical, emotional, social, intellectual, spiritual, occupational). We construct a high-quality Chinese dataset with span-level annotations and a domain-expert-informed annotation protocol. Our interpretable modeling approach integrates traditional machine learning with Transformer-based architectures, evaluated via 10-fold cross-validation and post-hoc interpretability techniques (LIME/SHAP). Contribution/Results: Our model achieves stable performance across all six dimensions (average F1-score = XX), significantly outperforming baselines. This work represents the first systematic operationalization of the six-dimensional wellness theory in social media analysis, providing a reproducible, regionally adaptable dataset and a transparent, interpretable technical pipeline for personalized mental health assessment and early intervention.

Technology Category

• Natural Language Processing: Interpretability, Analysis, and Evaluation of NLP Models
• Application Domains: Humanities & Computational Social Science
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Social Networks and Social Media: Computational social science

Abstract

We introduce a dataset for classifying wellness dimensions in social media user posts, covering six key aspects: physical, emotional, social, intellectual, spiritual, and vocational. The dataset is designed to capture these dimensions in user-generated content, with a comprehensive annotation framework developed under the guidance of domain experts. This framework allows for the classification of text spans into the appropriate wellness categories. We evaluate both traditional machine learning models and advanced transformer-based models for this multi-class classification task, with performance assessed using precision, recall, and F1-score, averaged over 10-fold cross-validation. Post-hoc explanations are applied to ensure the transparency and interpretability of model decisions. The proposed dataset contributes to region-specific wellness assessments in social media and paves the way for personalized well-being evaluations and early intervention strategies in mental health. We adhere to ethical considerations for constructing and releasing our experiments and dataset publicly on Github.

Problem

• Classify wellness dimensions in social media posts
• Evaluate models for multi-class wellness categorization
• Enable personalized well-being assessments and interventions

Innovation

• Dataset for holistic wellness dimensions analysis
• Traditional and transformer-based models evaluated
• Post-hoc explanations for model transparency