Explainable Deep Attention Active Learning for Sentimental Analytics of Mental Disorder
Bibliographic record
Abstract
With the increasing use of online mediums, Internet-delivered psychological treatments (IDPs) are becoming an essential tool for improving mental disorders. Online-based health therapies can help a large segment of the population with little resource investment. The task is greatly complicated by the overlapping emotions for specific mental health. Early adoption of a deep learning system presented severe difficulties, including ethical and legal considerations that contributed to a lack of trust. Modern models required highly interpretable, intuitive explanations that humans could understand. To achieve this, we present a deep attention model based on fuzzy classification that uses the linguistic features of patient texts to build emotional lexicons. In medical applications, a diversified dataset generates work. Active learning techniques are used to extend fuzzy rules and the learned dataset gradually. From this, the model can gain a reduction in labeling efforts in mental health applications. In this way, difficulties such as the amount of vocabulary per class, method of generation, the source of data, and the baseline for human performance level can be solved. Moreover, this work illustrates fuzzy explainability by using weighted terms. The proposed method incorporates a subset of unstructured data into the set for training and uses a similarity-based approach. The approach then updates the model training using the new training points in the subsequent cycle of the active learning mechanism. The cycle is repeated until the optimal solution is found. At this point, all unlabeled text is converted into the set for training. The experimental results show that the emotion-based enhancement improves test accuracy and helps develop quality criteria. In the blind test, the bidirectional LSTM architecture with an attention mechanism and fuzzy classification achieved an F1 score of 0.89.
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 | 0.001 | 0.002 |
| 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 | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 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 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".