Emotion-Based Reinforcement Attention Network for Depression Detection on Social Media: Algorithm Development and Validation
Bibliographic record
Abstract
BACKGROUND: Depression detection has recently received attention in the field of natural language processing. The task aims to detect users with depression based on their historical posts on social media. However, existing studies in this area use the entire historical posts of the users and select depression indicator posts. Moreover, these methods fail to effectively extract deep emotional semantic features or simply concatenate emotional representation. To solve this problem, we propose a model to extract deep emotional semantic features and select depression indicator posts based on the emotional states. OBJECTIVE: This study aims to develop an emotion-based reinforcement attention network for depression detection of users on social media. METHODS: The proposed model is composed of 2 components: the emotion extraction network, which is used to capture deep emotional semantic information, and the reinforcement learning (RL) attention network, which is used to select depression indicator posts based on the emotional states. Finally, we concatenated the output of these 2 parts and send them to the classification layer for depression detection. RESULTS: Experimental results of our model on the multimodal depression data set outperform the state-of-the-art baselines. Specifically, the proposed model achieved accuracy, precision, recall, and F1-score of 90.6%, 91.2%, 89.7%, and 90.4%, respectively. CONCLUSIONS: The proposed model utilizes historical posts of users to effectively identify users' depression tendencies. The experimental results show that the emotion extraction network and the RL selection layer based on emotional states can effectively improve the accuracy of detection. In addition, sentence-level attention layer can capture core posts.
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.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".