Explainable Multi-Emotion Mental Health Classification from Twitter Emotions Dataset Using Bidirectional GRU with LIME

References

[1] M. Mondal, S. Das, and A. Ghosh, “Emotion Detection from Social Media Posts,” arXiv preprint, arXiv:2302.05610, 2023.

[2] D. Ji, P. Zhou, and Y. Zhang, “MentalBERT: Publicly Available Pre-trained Language Models for Mental Healthcare,” in Proc. Int. Conf. Language Resources and Evaluation (LREC), Marseille, France, 2022, pp. 4382– 4390.

[3] D. Demszky, D. Movshovitz-Attias, J. Ko, A. Cowen, G. Nemade, and S. Ravi, “GoEmotions: A Dataset of Fine-Grained Emotions,” in Proc. 58th Annual Meeting of the Association for Computational Linguistics (ACL), Online, 2020, pp. 4040–4054.

[4] V. Vajre, M. Nair, and S. K. Singh, “PsychBERT: A Mental Health Lan-guage Model for Social Media,” in Proc. IEEE Int. Conf. Bioinformatics and Biomedicine (BIBM), Houston, TX, USA, 2021, pp. 1077– 1082.

[5] A. S. Abdelghafoor, R. Alghamdi, and H. Aljohani, “Natural Language Processing for Mental Health Interventions: A Systematic Review,” Translational Psychiatry, vol. 13, no. 309, pp. 1–12, 2023.

[6] M. Nourani, J. T. Hancock, and V. S. Subrahmanian, “Machine Learning and NLP in Mental Health: A Systematic Review,” Journal of Medical Internet Research, vol. 23, no. 5, e15708, 2021.

[7] R. Calvo, D. Milne, and S. Razavi, “Natural Language Processing for Mental Health Interventions,” Frontiers in Digital Health, vol. 3, art. no. 646532, 2021.

[8] S. Chancellor, E. P. S. Baumer, and M. De Choudhury, “Who is the ‘Human’ in Human- Centered Machine Learning? The Case of Predicting Mental Health from Social Media,” Proc. ACM on Human-Computer Interaction, vol. 3, no. CSCW, art. no. 152, pp. 1–32, 2019.

[9] Y. Rao, J. Li, and H. Xie, “Emotion Recognition of Social Media Users Based on Deep Learning,” Multimedia Tools and Applications, vol. 82, no. 2, pp. 2345–2364, 2023.

[10] J. Eisenberg, A. L. Beatty, and C. M. Lee, “Mixed Emotion Extraction from Social Media Text,” Data & Knowledge Engineering, vol. 149, art. no. 102259, 2024.

[11] M. T. Islam, A. K. Das, and M. S. Hossain, “Emotion Detection from Social Media Text Using Deep Learning,” Int. J. Advanced Computer Science and Applications, vol. 14, no. 5, pp. 312–320, 2023.

[12] A. H. Khan, S. U. Khan, and F. U. M. Abbasi, “Sentiment Analysis for Mental Health Monitoring Using Hybrid Deep Learning,” IEEE Access, vol. 10, pp. 12345–12356, 2022.

[13] S. Poria, N. Majumder, D. Hazarika, and R. Mihalcea, “Multimodal Emotion Recognition Using Deep Learning,” Neural Computing and Applications, vol. 35, no. 1, pp. 123–140, 2023.

[14] X. Glorot, A. Bordes, and Y. Bengio, “Deep Sparse Rectifier Neural Networks,” in Proc. 14th Int. Conf. Artificial Intelligence and Statistics (AISTATS), Fort Lauderdale, FL, USA, 2011, pp. 315–323.

[15] M. T. Ribeiro, S. Singh, and C. Guestrin, “Why Should I Trust You? Explaining Predictions of Any Classifier,” in Proc. 22nd ACM SIGKDD Int. Conf. Knowledge Discovery and Data Mining, San Francisco, CA, USA, 2016, pp. 1135–1144.

[16] F. Pedregosa et al., “Scikit-learn: Machine Learning in Python,” Journal of Machine Learning Research, vol. 12, pp. 2825–2830, 2011.

[17] A. Aldayel and M. Magdy, “Your Stance is Exposed! Analysing Possible Factors for Stance Detection on Social Media,” Proc. ACM on Human-Computer Interaction, vol. 3, no. CSCW, art. no. 210, pp. 1–20, 2019.

[18] U. Naseem, I. Razzak, and P. W. Eklund, “A Survey of Deep Learning for Mental Health Detection on Social Media,” IEEE Trans. Computational Social Systems, vol. 9, no. 5, pp. 1322–1340, 2022.

[19] S. M. Mohammad and P. D. Turney, “Crowdsourcing a Word–Emotion Association Lexicon,” Computational Intelligence, vol. 29, no. 3, pp. 436–465, 2013.

[20] L. H. Lee, C. H. Chen, and S. M. Lee, “Explainable Transformer-Based Model for Emotion Detection on Social Media,” IEEE Access, vol. 11, pp. 24567–24580, 2023.

[21] T. Chen and C. Guestrin, “XGBoost: A Scalable Tree Boosting System,” in Proc. 22nd ACM SIGKDD Int. Conf. Knowledge Discovery and Data Mining, San Francisco, CA, USA, 2016, pp. 785–794.

[22] S. Hochreiter and J. Schmidhuber, “Long Short-Term Memory,” Neural Computation, vol. 9, no. 8, pp. 1735–1780, 1997.

[23] K. Cho, B. van Merrie¨nboer, C. Gulcehre, D. Bahdanau, F. Bougares, H. Schwenk, and Y. Bengio, “Learning Phrase Representations Using RNN Encoder–Decoder for Statistical Machine Translation,” in Proc. Conf. Empirical Methods in Natural Language Processing (EMNLP), Doha, Qatar, 2014, pp. 1724–1734.

[24] D. Bahdanau, K. Cho, and Y. Bengio, “Neural Machine Transla-tion by Jointly Learning to Align and Translate,” arXiv preprint, arXiv:1409.0473, 2014.

[25] S. M. Mohammad, F. Bravo-Marquez, M. Salameh, and S. Kiritchenko, “SemEval-2018 Task 1: Affect in Tweets,” in Proc. 12th Int. Workshop on Semantic Evaluation (SemEval), New Orleans, LA, USA, 2018, pp. 1–17.

@article{1718014,
      author = {Awaz Bhujel, Dr. M. N. Nachappa},
      title = {Explainable Multi-Emotion Mental Health Classification from Twitter Emotions Dataset Using Bidirectional GRU with LIME},
      journal = {Iconic Research And Engineering Journals},
      year = {2026},
      volume = {9},
      number = {11},
      pages = {2745-2756},
      issn = {2456-8880},
      url = {https://www.irejournals.com/formatedpaper/1718014.pdf},
      abstract = {Social media platforms like X (formerly Twitter) have become common places where people share their feelings. These emotional expressions can sometimes reveal early signs of mental health conditions. However, detecting multiple emotions at the same time from short, casual social media posts is still a difficult task. Most previous studies either focus only on positive and negative sentiment or try to identify just one emotion per post. This leaves a clear gap in systems that can recognize several emotions together and explain how they reached their decisions. In this paper, we propose an explainable deep learning framework for multi-emotion mental health classification using the Twitter Emotions Dataset. We built a balanced dataset of 60,000 tweets covering six emotion categories: anger, fear, joy, love, sadness, and surprise. We then compared five deep learning models (LSTM, GRU, BiLSTM, BiGRU, and BiLSTM with Bahdanau attention) against three traditional machine learning methods (Logistic Regression, Naive Bayes, and XGBoost). Among all models, the Bidirectional GRU (BiGRU) achieved the best classification performance. It proved particularly good at capturing context from both directions in short text. To meet the need for interpretability in mental health applications, we applied Local Interpretable Model-agnostic Explanations (LIME) to the BiGRU model. LIME produced word-level visualizations showing which words contributed most to each of the six emotion predic-tions. These explanations make the model’s behavior transparent and support trustworthy AI in mental health monitoring.},
      keywords = {Explainable AI (XAI), Mental Health, Multi-Emotion Classification, BiGRU, LIME, Twitter Dataset, Deep Learning, NLP, Sentiment Analysis},
      month = {May},
      doi = {https://doi.org/10.64388/IREV9I11-1718014}
  }