An attention-based hybrid architecture with explainability for depressive social media text detection in Bangla
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Mental health has become a major concern in recent years. Social media have been increasingly used as platforms to gain insight into a person’s mental health condition by analysing the posts and comments, which are textual in nature. By analysing these texts, depressive posts can be detected. To facilitate this process, this work presents an attention-based bidirectional Long Short-Term Memory (LSTM)- Convolutional Neural Network (CNN) based model to detect depressive Bangla social media texts, which is lighter and more robust than the conventional models and provides better performance. A dataset containing such Bangla texts was also developed in this work to mitigate the scarcity. Different preprocessing stages were followed, and three embeddings were used in this task. Thanks to the attention mechanism, the proposed model achieved an accuracy of 94.3% with 92.63% of sensitivity and 95.12% of specificity. When tested on other languages, such as English, the proposed model performed remarkably. The robustness and explainability of the proposed model were also discussed in this paper. Additionally, when compared with classical machine learning models, ensemble approaches, transformers, other similar models, and existing architectures, the proposed model outperformed them.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it