Elamur Research⌕
☰
home / papers / 2026 / Depression Detection in Modern Greek

Depression Detection in Modern Greek

Paper · 2026 · Vivian Stamou, George Mikros, George Markopoulos et al.

Summary

Most NLP work on depression screening covers English, so this study builds a first benchmark for Modern Greek social media. It reuses the Greek depression corpus of Stamou et al. (2024). That corpus contains tweet histories of 51 users who said publicly that they had been diagnosed with depression. It is paired with two control sets: randomly sampled users (CC1) and a topic-matched group (CC2) meant to reduce thematic bias. The authors test classical classifiers built in PyCaret, using either TF-IDF or LIWC features, against four fine-tuned transformers: GreekBERT, GreekSocialBERT, mBERT and XLM-R. They also run a stress test that pairs one confirmed depressed user with a closely matched control. Transformers perform best, with F1 around 0.95, but they are hard to interpret. To address this, the authors apply SHAP to LIWC features with an Extra Trees model. The resulting cues match known clinical signs, such as less talk about work, social withdrawal and anhedonia-like loss of motivation. The work gives a reference point for Greek and argues for screening tools grounded in interpretable evidence.

Key points

  • Data: The Depression Corpus has 51 self-reporting users and 659,189 tweets. Modelling used a balanced 10K-tweet sample, with 5,000 depression tweets and 5,000 from CC2. Data were split by user (80/20, 5-fold CV), so no user appears in both training and test sets.
  • Training size: Scaling from 10K to 40K tweets raised F1 only slightly, from 0.9398 to 0.9582.
  • LIWC features on CC2: LightGBM was the strongest model (F1 = 0.7123, accuracy 0.7170). Extra Trees reached F1 = 0.6779 and was kept for interpretability analysis.
  • TF-IDF features: These (1–3-grams, 5,000 features) clearly beat LIWC. Ridge and LDA exceeded 0.87 accuracy.
  • Transformers: The models were bert-base-greek-uncased-v1, greeksocialbert-base-v2, bert-base-multilingual-cased and xlm-roberta-base. Each underwent 4 epochs of fine-tuning using a 1e-5 step size and batch size 16, and reached F1 ≈ 0.95.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓Hamad bin Khalifa UniversityDoha · Qatar141 works · 14,175 citations🎓National and Kapodistrian University of AthensAthens · Greece2 works · 74 citations

 Fields, topics and concepts

📚 Shelves 5The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksPsychology · 1,552 worksExperimental and Cognitive Psychology · 870 worksSocial Psychology · 306 works

Cite

Cite as Vivian Stamou, George Mikros, George Markopoulos and Spyridoula Varlokosta. 2026. Depression Detection in Modern Greek. Proceedings of the Sixth Resources and ProcessIng of linguistic, para-linguistic and extra-linguistic Data from people with various forms of cognitive/psychiatric/developmental impairments in cooperation with the MENTAL.ai consortium.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.63317/2gdmrug2fvmw

@inproceedings{stamou2026depression,
  title = {{Depression Detection in Modern Greek}},
  author = {Vivian Stamou, George Mikros, George Markopoulos and Spyridoula Varlokosta},
  booktitle = {Proceedings of the Sixth Resources and ProcessIng of linguistic, para-linguistic and extra-linguistic Data from people with various forms of cognitive/psychiatric/developmental impairments in cooperation with the MENTAL.ai consortium},
  year = {2026},
  doi = {10.63317/2gdmrug2fvmw},
  url = {https://doi.org/10.63317/2gdmrug2fvmw}
}

References

26 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Moez Ali. 2020. PyCaret: An open source, low-code machine learning library in Python. PyCaret version 1.0.0. find
  2. Salma Almouzini, Maher khemakhem, and Asem Alageel. 2019. Detecting arabic depressed users from twitter data. Procedia Comput. Sci., 163(C):257–265. find
  3. Meysam Asgari, Jeffrey Kaye, and Hiroko Dodge. 2017. Predicting mild cognitive impairment from spontaneous spoken utterances. Alzheimer’s & Dementia: Translational Research & Clinical Interventions, 3(2):219–228. find
  4. Deanna M. Barch, David Pagliaccio, and Katherine R. Luking. 2016. Mechanisms underlying motivational deficits in psychopathology: similarities and differences in depression and schizophrenia. Current Topics in Behavioral Neurosciences, 27:411–449. find
  5. Ana-Maria Bucur, Marcos Zampieri, Tharindu Ranasinghe, and Fabio Crestani. 2025. A survey on multilingual mental disorders detection from social media data. CoRR, abs/2505.15556. find
  6. Stevie Chancellor and Munmun De Choudhury. 2020. Methods in predictive techniques for mental health status on social media: a critical review. NPJ Digital Medicine, 3. find
  7. Glen Coppersmith, Mark Dredze, Craig Harman, and Kristy Hollingshead. 2015. From ADHD to SAD: Analyzing the language of mental health on Twitter through self-reported diagnoses. In Proceedings of the 2nd Workshop on Computational Linguistics and Clinical Psychology: From Linguistic Signal to Clinica… find
  8. Munmun De Choudhury, Michael Gamon, Scott Counts, and Eric Horvitz. 2013. Predicting depression via social media. In Proceedings of ICWSM, pages 128–137. find
  9. Johannes C. Eichstaedt, Robert J. Smith, Raina M. Merchant, Lyle H. Ungar, Patrick Crutchley, Daniel Preotiuc-Pietro, David A. Asch, and H. Andrew Schwartz. 2018. Facebook language predicts depression in medical records. Proceedings of the National Academy of Sciences, 115:11203–11208. find
  10. Lucía Gómez-Zaragozá, Javier Marín-Morales, Mariano Alcañiz, and Mohammad Soleymani. 2025. Speech and text foundation models for depression detection: Cross-task and cross-language evaluation. find
  11. Keith Harrigian, Carlos Aguirre, and Mark Dredze. 2021. On the state of social media data for mental health research. In Proceedings of the Seventh Workshop on Computational Linguistics and Clinical Psychology: Improving Access, pages 15–24, Online. Association for Computational Linguistics. find
  12. Ahmed Husseini Orabi, Prasadith Buddhitha, Mahmoud Husseini Orabi, and Diana Inkpen. 2018. Deep learning for depression detection of Twitter users. In Proceedings of the Fifth Workshop on Computational Linguistics and Clinical Psychology: From Keyboard to Clinic, pages 88–97, New Orleans, LA. Associ… find
  13. Glorianna Jagfeld, Fiona Lobban, Paul Rayson, and Steven Jones. 2021. Understanding who uses Reddit: Profiling individuals with a self-reported bipolar disorder diagnosis. In Proceedings of the Seventh Workshop on Computational Linguistics and Clinical Psychology: Improving Access, pages 1–14, Onlin… find
  14. Zunaira Jamil, Diana Inkpen, Prasadith Buddhitha, and Kenton White. 2017. Monitoring tweets for depression to detect at-risk users. In Proceedings of the Fourth Workshop on Computational Linguistics and Clinical Psychology — From Linguistic Signal to Clinical Reality, pages 32–40, Vancouver, BC. Ass… find
  15. Danxia Liu, Xing Lin Feng, Farooq Ahmed, Muhammad Shahid, and Jing Guo. 2022. Detecting and measuring depression on social media using a machine learning approach: Systematic review. JMIR Ment Health, 9(3):e27244. find
  16. James W. Pennebaker, Roger J. Booth, Ryan L. Boyd, and Martha E. Francis. 2015. Linguistic Inquiry and Word Count: LIWC2015. Pennebaker Conglomerates, Austin, TX. find
  17. Peng Qi, Yuhao Zhang, Yuhui Zhang, Jason Bolton, and Christopher D. Manning. 2020. Stanza: A python natural language processing toolkit for many human languages. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics: System Demonstrations, pages 101–108, Online.…
  18. Philip Resnik, Anderson Garron, and Rebecca Resnik. 2013. Using topic modeling to improve prediction of neuroticism and depression in college students. In Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, pages 1348–1353, Seattle, Washington, USA. Association fo… find
  19. Zhiling Zhang, Siyuan Chen, Mengyue Wu, and Kenny Zhu. 2022. Symptom identification for interpretable detection of multiple mental disorders on social media. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 9970–9985, Abu Dhabi, United Arab Emirates. A… find
  20. Wesley Ramos dos Santos, Rafael Lage de Oliveira, and Ivandré Paraboni. 2023. Setembrobr: a social media corpus for depression and anxiety disorder prediction. Lang. Resour. Eval., 58(1):273–300. find
  21. Vivian Stamou, George Mikros, George Markopoulos, and Spyridoula Varlokosta. 2024. Establishing control corpora for depression detection in Modern Greek: Methodological insights. In Proceedings of the Fifth Workshop on Resources and ProcessIng of linguistic, para-linguistic and extra-linguistic Data… find
  22. Michael M. Tadesse, Hongfei Lin, Bo Xu, and Liang Yang. 2019. Detection of depression-related posts in reddit social media forum. IEEE Access, 7:44883–44893. find
  23. Michael T. Treadway and David H. Zald. 2011. Reconsidering anhedonia in depression: lessons from translational neuroscience. Neuroscience and biobehavioral reviews, 35 3:537–55. find
  24. Rebecca Watson, Kate Harvey, Ciara McCabe, and Shirley Reynolds. 2020. Understanding anhedonia: A qualitative study exploring loss of interest and pleasure in adolescent depression. European Child & Adolescent Psychiatry, 29(4):489– 499. find
  25. Sourabh Zanwar, Daniel Wiechmann, Yu Qiao, and Elma Kerz. 2023. SMHD-GER: A large-scale benchmark dataset for automatic mental health detection from social media in German. In Findings of the Association for Computational Linguistics: EACL 2023, pages 1526–1541, Dubrovnik, Croatia. Association for C… find
  26. Zhenwen Zhang, Jianghong Zhu, Zhihua Guo, Yu Zhang, Zepeng Li, and Bin Hu. 2024. Natural find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/stamou-2026-depression-detection-modern-greek/