Saved in:
| Main Author: | |
|---|---|
| Format: | Recurso digital |
| Language: | English |
| Published: |
Zenodo
2025
|
| Subjects: | |
| Online Access: | https://doi.org/10.5281/zenodo.17140286 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Table of Contents:
- <p>This report examines the use of artificial intelligence (AI) for the early detection of mental health concerns expressed through social media. As online platforms increasingly serve as outlets for emotional expression, they offer a valuable yet complex space for identifying psychological distress, including depression, anxiety, and suicidal ideation.</p> <p>The study presents <strong>MentalBERT</strong>, a domain-adapted transformer model fine-tuned on Reddit data and evaluated on both Reddit and Twitter datasets. MentalBERT demonstrated a notable improvement over classical machine learning baselines, achieving a macro F1 score of 0.76 and exhibiting strong sensitivity to nuanced and euphemistic language associated with distress.</p> <p>Ethical integrity was embedded throughout the research process. Measures included data anonymisation, fairness-aware calibration, explainability mechanisms, and safeguards against algorithmic bias. External validation by Adil Majeed, a data scientist at Intellytics, affirmed both the model’s technical robustness and its ethical suitability for deployment in sensitive contexts.</p> <p>This work contributes a scalable, context-aware, and ethically grounded framework for digital mental health detection. It advocates for the responsible development of AI systems that emphasise interpretability, inclusivity, and respect for user dignity. Particularly within preventative care and clinical triage settings.</p>