Mental Health Prediction Using Machine Learning and Explainable AI: A Review of Text-Based and Emotion-Aware Approaches

This review examined machine learning and Transformer-based approaches for text-based mental health prediction. The literature highlights the importance of multiclass prediction, systematic cross-model benchmarking, emotion feature integration, and model explainability. Based on the identified gaps, a framework was proposed that compares classical ML and Transformer models, integrates emotion features with textual representations, and applies Explainable AI techniques such as SHAP and LIME. The framework provides a structured direction for developing more transparent and interpretable mental health prediction systems from textual data.

[1] M. M. Tadesse, H. Lin, B. Xu, and L. Yang, “Detection of depression-related posts in Reddit social media forum,” IEEE Access, vol. 7, pp. 44883–44893, 2019, doi: 10.1109/ACCESS.2019.2909180. [2] R. Haque, N. Islam, M. Islam, and M. M. Ahsan, “A comparative analysis on suicidal ideation detection using NLP, machine, and deep learning,” Technologies, vol. 10, no. 3, Art. no. 57, 2022, doi: 10.3390/technologies10030057 [3] T. Sood, P. Garg, H. Pannu, and S. Dullat, “Machine learning algorithms and NLP for anxiety and depression detection using Twitter data,” in Proc. 2025 Int. Conf. Cognitive Computing in Engineering, Communications Sciences and Biomedical Health Informatics (IC3ECSBHI), 2025, pp. 450–455, doi: 10.1109/IC3ECSBHI63591.2025.10990488. [4] İ. Baydili, B. Tasci, and G. Tasci, “Deep learning-based detection of depression and suicidal tendencies in social media data with feature selection,” Behavioral Sciences, vol. 15, no. 3, Art. no. 352, 2025, doi: 10.3390/bs15030352. [5] A. Nadeem, M. Naveed, and M. I. Satti, “Depression detection based on hybrid deep learning SSCL framework using self-attention mechanism: An application to social networking data,” Sensors, vol. 22, no. 24, Art. no. 9775, 2022, doi: 10.3390/s22249775. [6] S. Zhou and M. Mohd, “Mental health safety and depression detection in social media text data: A classification approach based on a deep learning model,” IEEE Access, vol. 13, pp. 63284–63297, 2025, doi: 10.1109/ACCESS.2025.3559170. [7] I. Arshad, S. Naseem, M. Mussarat, A. Zanib, S. Al-Otaibi, A. Yousif, T. Mahmood, and A. R. Khan, “A multi-platform approach to optimizing depression detection in social media through feature engineering and contextual analysis,” International Journal of Computational Intelligence Systems, vol. 19, Art. no. 97, 2026, doi: 10.1007/s44196-026-01225-y. [8] B. G. Bokolo and Q. Liu, “Advanced comparative analysis of machine learning and Transformer models for depression and suicide detection in social media texts,” Electronics, vol. 13, no. 20, Art. no. 3980, 2024, doi: 10.3390/electronics13203980. [9] B. G. Bokolo and Q. Liu, “Deep learning-based depression detection from social media: Comparative evaluation of ML and Transformer techniques,” Electronics, vol. 12, no. 21, Art. no. 4396, 2023, doi: 10.3390/electronics12214396. [10] M. A. Abbas, K. Munir, A. Raza, N. Abdel Samee, M. M. Jamjoom, and Z. Ullah, “Novel Transformer based contextualized embedding and probabilistic features for depression detection from social media,” IEEE Access, vol. 12, pp. 54087–54100, 2024, doi: 10.1109/ACCESS.2024.3387695.