Cognitive-Linguistic Indicators of Depression in Online Communities: Analysed by DistilBERT and Holographic Reduced Representation
Quick Take
This study demonstrates that a hybrid model combining DistilBERT embeddings with Holographic Reduced Representation significantly enhances depression detection in online text, achieving a macro F1 score of 0.94 compared to 0.80 for the TF-IDF baseline. The model utilizes cognitive distortions from Beck's theory and shows improved performance metrics such as F1 and AUC through a robust classification pipeline.
Key Points
- Hybrid model combines DistilBERT and Holographic Reduced Representation for better results.
- Achieved macro F1 score of 0.94, outperforming TF-IDF baseline score of 0.80.
- 5-fold cross-validation improved F1 from 0.83 to 0.92 and AUC from 0.958 to 0.981.
- Cognitive distortions like pronoun density and absolutist words were key features.
- Study utilized Reddit posts from depression-related and control communities.
Article Excerpt
From source RSS / original summaryarXiv:2606. 00026v1 Announce Type: new Abstract: This paper investigates whether combining cognitively grounded linguistic features with transformer-based embeddings improves automated detection of depression in online text. Using Beck's Cognitive Theory of Depression, the study extracts cognitive distortions as measurable features, including first-person pronoun density, absolutist words, and negative emotion in Reddit posts from depression-related and control communities.
Using a subset of the Kaggle Reddit Suicide and Depression Detection dataset, two classification pipelines are compared, a TF-IDF embedding with Naive Bayes as a baseline, and a hybrid model that concatenates DistilBERT sentence embeddings with Holographic Reduced Representation (HRR) vectors encoding the cognitive-linguistic features, followed by Logistic Regression. The hybrid DistilBERT HRR model achieves a macro F1 score of 0. 94 versus 0.
80 for the TD-IDF baseline, with 5-fold cross validation F1 improving from 0. 83 to 0. 92, and AUC from 0. 958 to 0. 981.
Reader Mode unavailable (could not extract clean content).
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CL
See more →Time to REFLECT: Can We Trust LLM Judges for Evidence-based Research Agents?
The REFLECT benchmark reveals that current LLM judges are unreliable, achieving below 55% accuracy in evaluating reasoning and evidence use, highlighting the need for improved evaluation methods for deep research agents.