跳到正文
原文
Nature 健康护理主题·· 2 天前AI 评分37

基于细调 BERT-GRU 框架的社交媒体心理健康情感分析

Mental health sentiment analysis on social media using a fine-tuned BERT–GRU framework

AI 导读

该研究提出了一种结合细调 BERT 和 GRU 的混合深度学习框架 FBert-GRU,用于从社交媒体数据中提取心理健康信号。实验结果显示,该方法在多个心理健康条件的检测上达到了96%的准确率、87%的精确率、82%的召回率、81%的F1分数和94%的AUC-ROC值,优于基准模型。该框架为大规模心理健康评估和行为趋势分析提供了有效策略。

正文

We’re sharing this article early to provide faster access to peer-reviewed, accepted research. It is citable and carries a permanent DOI. This version is subject to further edits and will be replaced automatically by the final Version of Record. All legal disclaimers apply.

Abstract

Mental health problems, such as depression, anxiety, stress, bipolar disorder, and thoughts of suicide are a serious health issue in the world today. Because users of social media platforms generate vast amounts of text, these data can provide researchers with insight into their emotional and psychological condition. However, owing to linguistic variability, contextual ambiguity, and the complex language of emotions, extracting meaningful indications of mental health from noisy, unstructured social media data is challenging. In this study, the authors propose a hybrid deep-learning framework, called FBert-GRU, which combines the fine-tuned Bidirectional Encoder Representations from Transformers (BERT) with Gated Recurrent Units (GRU) to learn the contextual semantic information and sequential emotional patterns. The framework was evaluated on a public mental health dataset comprising social media posts that were labeled for several mental health conditions. The experimental results showed an accuracy of 96%, a precision of 87%, a recall of 82%, an F1-score of 81% and an AUC-ROC of 94%. The experimental results showed that the proposed method using contextual embeddings and sequential learning improved the detection of mental health-related signals from social media text compared with the benchmark models. The suggested framework offers a promising strategy for conducting mental health evaluations and behavioral trend analysis at a large scale and for the digital monitoring of mental health.

Abbreviations

BERT:

Bidirectional encoder representations from transformers

GRU:

Gated recurrent unit

FBert-GRU:

Fine-tuned BERT with gated recurrent unit

DNN:

Deep neural network

CNN:

Convolutional neural network

RNN:

Recurrent neural network

LSTM:

Long short-term memory

BiLSTM:

Bidirectional long short-term memory

NLP:

Natural language processing

ABSA:

Aspect-based sentiment analysis

MLP:

Multilayer perceptron

SVM:

Support vector machine

OCR:

Optical character recognition

NPV:

Negative predictive value

ROC:

Receiver operating characteristic

AUC-ROC:

Area under the receiver operating characteristic curve

SHAP:

SHapley Additive exPlanations

SOTA:

State-of-the-art

Acknowledgements

The authors gratefully acknowledge Manipal University Jaipur, Jaipur for providing research facilities and assistance.

Funding

Open access funding provided by Manipal University Jaipur.

Author information

Author notes

  1. Dheeraj Tandon

    Present address: Institute of Engineering and Technology, Sitapur, Uttar Pradesh, Sitapur, India

Authors and Affiliations

  1. School of Computing Science and Engineering, Galgotias University, Greater Noida, Uttar Pradesh, India

    Sudheer Kumar Singh

  2. Department of Computer Science, ERA University, Lucknow, India

    Nagendra Kumar Singh & Santosh Kumar

  3. School of Humanities and Social Sciences, Sharda University, Greater Noida, Uttar Pradesh, India

    Ranjana Singh

  4. Department of Computer Applications, Manipal University Jaipur, Jaipur, India

    Virendra Singh

Authors

  1. Sudheer Kumar Singh
  2. Nagendra Kumar Singh
  3. Dheeraj Tandon
  4. Santosh Kumar
  5. Ranjana Singh
  6. Virendra Singh

Corresponding author

Correspondence to Virendra Singh.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Singh, S.K., Singh, N.K., Tandon, D. et al. Mental health sentiment analysis on social media using a fine-tuned BERT–GRU framework. Sci Rep (2026). https://doi.org/10.1038/s41598-026-75415-1

Download citation

  • Received: 17 June 2026

  • Accepted: 06 October 2026

  • Published: 18 November 2026

  • DOI: https://doi.org/10.1038/s41598-026-75415-1

Keywords

来源:Nature 健康护理主题 · nature.com