Utilizing Deep Learning Techniques for Mental Disorder Prediction and Support Using Reddit Data
Bibliographic record
Abstract
The increasing prevalence of mental health disorders necessitates innovative approaches to early detection and intervention. Social media platforms like Reddit provide a rich source of textual data that can be leveraged to identify mental health issues based on user-generated content. This study aims to classify mental health disorders by analyzing Reddit comments. The dataset used is Reddit SuicideWatch and Mental Health Collection that includes 54,412 posts, which are classified into several mental health disorders. After that, preprocessing steps were done, which includes tokenization, padding, word embeddings, and handling imbalance data. Then, a Custom-CNN (Convolutional Neural Network) and a Custom- RNN (Recurrent Neural Network) were used to classify the data. These evaluation metrics were then used to evaluate the models: accuracy, precision, recall, F1-score, and ROC AUC Score. Notable results are that the RNN has a higher accuracy of 0.65 compared to 0.64 by CNN. In addition, both RNN and CNN also have high ROC AUC scores of 0.88 and 0.87 respectively. In conclusion, the results show that the models are good at differentiating one class from another, however, both models have problems accurately classifying the correct mental health disorder from the texts. This shows promise as the models can definitely be improved with better and more balanced data, and exploration on LLMs (Large Language Models) can also be valuable for this study and issue. In addition to that, those improvements can be applied to the act of intervention by integrating it to technology such as a conversational AI or wearable actuators that have the ability to predict and prevent through real-time monitoring.
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 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 | 0.002 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.001 |
| 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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, 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".