Transfer Learning for Risk Classification of Social Media Posts: Model\n Evaluation Study
Bibliographic record
Abstract
Mental illness affects a significant portion of the worldwide population.\nOnline mental health forums can provide a supportive environment for those\nafflicted and also generate a large amount of data which can be mined to\npredict mental health states using machine learning methods. We benchmark\nmultiple methods of text feature representation for social media posts and\ncompare their downstream use with automated machine learning (AutoML) tools to\ntriage content for moderator attention. We used 1588 labeled posts from the\nCLPsych 2017 shared task collected from the Reachout.com forum (Milne et al.,\n2019). Posts were represented using lexicon based tools including VADER,\nEmpath, LIWC and also used pre-trained artificial neural network models\nincluding DeepMoji, Universal Sentence Encoder, and GPT-1. We used TPOT and\nauto-sklearn as AutoML tools to generate classifiers to triage the posts. The\ntop-performing system used features derived from the GPT-1 model, which was\nfinetuned on over 150,000 unlabeled posts from Reachout.com. Our top system had\na macro averaged F1 score of 0.572, providing a new state-of-the-art result on\nthe CLPsych 2017 task. This was achieved without additional information from\nmeta-data or preceding posts. Error analyses revealed that this top system\noften misses expressions of hopelessness. We additionally present\nvisualizations that aid understanding of the learned classifiers. We show that\ntransfer learning is an effective strategy for predicting risk with relatively\nlittle labeled data. We note that finetuning of pretrained language models\nprovides further gains when large amounts of unlabeled text is available.\n
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 | 0.006 | 0.010 |
| Meta-epidemiology (narrow) | 0.002 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.002 | 0.003 |
| Insufficient payload (model declined to judge) | 0.005 | 0.003 |
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".