图书情报工作

图书情报工作 >

2026 , Vol. 70 >Issue 15: 49 - 59

DOI: https://doi.org/10.13266/j.issn.0252-3116.2026.15.004

研究论文

融合语义增强的社交媒体健康信息可信度细粒度评估

  • 刘成山 ,
  • 唐璇 ,
  • 秦春秀
展开
  • 西安电子科技大学经济与管理学院 西安 710126
刘成山,副研究员,硕士生导师;唐璇,硕士研究生,通信作者,E-mail:23061212962@stu.xidian.edu.cn;秦春秀,教授,博士,硕士生导师。

收稿日期: 2025-05-22

  修回日期: 2025-09-02

  网络出版日期: 2026-07-30

收起

Fine-Grained Credibility Assessment of Social Media Health Information with Semantic Enhancement

  • Liu Chengshan ,
  • Tang Xuan ,
  • Qin Chunxiu
Expand
  • School of Economics and Management, Xidian University, Xi'an 710126
Liu Chengshan, Associate Researcher, Master's Supervisor; Tang Xuan, Master's Candidate, Corresponding Author, Email:23061212962@stu.xidian.deu.cn; Qin Chunxiu, Professor, PhD, Master's Supervisor.

Received date: 2025-05-22

  Revised date: 2025-09-02

  Online published: 2026-07-30

Fold

摘要

[目的/意义] 以期解决社交媒体健康信息标注数据稀缺导致模型泛化能力弱、短文本语义稀疏、分类粒度较粗的问题，探究增强健康信息语义特征方法、细化可信度分类，提高健康信息可信度评估的可解释性与准确性。[方法/过程] 首先，构建BERT-CNN-BiLSTM语义增强架构，引入依存句法分析获取候选关系元组，融合外部医学知识库生成标准关系元组；其次，融合评论语义、情感特征等维度构建可信度评估指标体系；最后依据可信度程度将健康信息划分为5类，实现健康信息细粒度评估。[结果/结论] 与专家标注的数据集和基线模型对比验证结果表明，提出的模型在五分类任务中准确率和F1值分别达到90.8%、89.0%，验证提出的融合语义增强的细粒度可信度评估模型能有效提高社交媒体健康信息的评估效果。

关键词： 语义增强; 信息可信度; 细粒度评估; 社交媒体

本文引用格式

刘成山 , 唐璇 , 秦春秀 . 融合语义增强的社交媒体健康信息可信度细粒度评估[J]. 图书情报工作, 2026 , 70(15) : 49 -59 . DOI: 10.13266/j.issn.0252-3116.2026.15.004

Abstract

[Purpose/Significance] This study aims to address the challenges of weak model generalization, semantic sparsity in short texts, and coarse classification granularity, which are caused by scarce annotated health information data on social media. It explores methods to enhance semantic features of health information, refine credibility classification, and improve the interpretability and accuracy of health information credibility assessment. [Method/Process] First, a BERT-CNN-BiLSTM semantic enhancement architecture was constructed. Dependency parsing was employed to extract candidate relation tuples, which were then fused with an external medical knowledge base to generate standardized relation tuples. Second, a credibility assessment metric system was established by integrating dimensions such as comment semantics and sentiment features. Finally, health information was categorized into five levels based on its credibility degree, enabling fine-grained assessment. [Result/Conclusion] Comparison with expert-annotated datasets and baseline models demonstrates that the proposed model achieves 90.8% accuracy and 89.0% F1 score on the five-class classification task. These results confirm that the proposed fusion semantic enhancement-based fine-grained credibility assessment model effectively improves the health information assessment on social media.

Key words： semantic enhancement; information credibility; fine-grained assessment; social media

参考文献

[1] 国务院办公厅. “ 十四五” 国民健康规划[R]. 国务院办公厅, 2022. (General office of the state council of the People’s Republic of China. China to promote national health during 14th five-year plan[R]. General Office of the State Council of the People’s Republic of China.)
[2] 邓胜利, 顾一飞. 网络虚假健康信息研究综述：认知、行为与治理[J]. 图书馆杂志, 2022, 41(5): 14-22. (DENG S L, GU Y F. A review of online health misinformation: recognition, action and governance[J]. Library journal, 2022, 41(5): 14-22.)
[3] 宋宁远, 裴雷, 王春迎. 科学论文语义增强的研究进展与趋势研判[J]. 图书情报工作, 2021, 65(1): 82-90. (SONG N Y, PEI L, WAGN C Y. The survey and tendency of semantic enrichment for scientific papers[J]. Library and information service, 2021, 65(1): 82-90.)
[4] 杨大伟, 徐西海, 宋威. 结合语义增强和感知注意力的关系抽取方法[J]. 计算机应用, 2025, 45(6): 1801-1808. (YANG D W, XU X H, SONG W. Relation extraction method combining semantic enhancement and perception attention[J]. Journal of computer applications, 2025, 45(6): 1801-1808.)
[5] 张梦芸, 丁敬达. 面向短文本分类的语义增强研究[J]. 图书情报工作, 2023, 67(9): 4-11, 3. (ZHANG M Y, DING J D. Research on semantic enhancement for short text classification[J]. Library and information service, 2023, 67(9): 4-11, 3.)
[6] 李云, 卢志翔, 刘姝伊, 等. 基于深度多模态关联学习的短视频多标签分类研究[J]. 数据分析与知识发现, 2024, 8(7): 77-88. (LI Y, LU Z X, LIU S Y, et al. Research on micro-video multi-label classification based on deep multimodal association learning[J]. Data analysis and knowledge discovery, 2024, 8(7): 77-88.)
[7] 王昊, 龚丽娟, 周泽聿, 等. 融合语义增强的社交媒体虚假信息检测方法研究[J]. 数据分析与知识发现, 2023, 7(2): 48- 60. (WANG H, GONG L J, ZHOU Z Y, et al. Detecting mis/disinformation from social media with semantic enhancement [J]. Data analysis and knowledge discovery, 2023, 7(2): 48-60.)
[8] 孙冉, 安璐. 突发公共卫生事件中谣言识别研究[J]. 情报资料工作, 2021, 42(5): 42-49. (SUN R, AN L. Research on rumor identification in public health emergency [J]. Information and documentation services, 2021, 42(5): 42-49.)
[9] 姚艺倍, 曹锦丹, 冯傅昭, 等. 国内外居民数字健康素养影响因素的meta分析[J]. 医学与社会, 2024, 37(11): 122-129. (YAO Y B, CAO J D, FENG F Z, et al. A meta-analysis of factors influencing digital health literacy among residents at home and abroad[J]. Medicine and society, 2024, 37(11): 122-129.)
[10] 宋士杰, 赵宇翔, 朱庆华. iField视域下的信息可信度研究: 概念溯源、主题演化与未来展望[J]. 中国图书馆学报, 2022, 48(1): 107-126. (SONG S J, ZHAO Y X, ZHU Q H. Information credibility research in the iField: conceptual development, topic evolution, and future direction[J]. Journal of library science in China, 2022, 48(1): 107-126.)
[11] YUNG S C, YAN Z, JACKE G. The effects of information source and eHealth literacy on consumer health information credibility evaluation behavior[J]. Computers in human behavior, 2021, 115:106629.
[12] EGALA S B, LIANG D C, BOATENG D. Social media healthrelated information credibility and reliability: an integrated user perceived quality assessment[J]. IEEE transactions on engineering management, 2024, 71: 5018-5029.
[13] SWARUP P, SANTOSH S R. An attention-based deep learning model for credibility assessment of online health information[J]. Computational intelligence, 2023, 39(5): 832-859.
[14] SHAFQAT W, BYUN Y C, PARK N. Effectiveness of machine learning approaches towards credibility assessment of crowdfunding projects for reliable recommendations[J]. Applied sciences, 2020, 10(24): 9062.
[15] 宋士杰, 赵宇翔, 宋小康, 等. 互联网环境下失真健康信息可信度判断的影响因素研究[J]. 中国图书馆学报, 2019, 45(4): 72-85. (SONG S J, ZHAO Y X, SONG X K, et al. Investigating the influential factors of consumer’s credibility judgment on health misinformation[J]. Journal of library science in China, 2019, 45(4): 72-85.)
[16] 徐孝婷, 张亭亭, 朱庆华. 在线健康社区中信息框架对HPV疫苗接种的影响研究—— 以信息可信度为中介变量[J]. 图书与情报, 2020(5): 39-47. (XU X T, ZHANG T T, ZHU Q H. Research on the impact of message framing on HPV vaccination in online health vommunities: a mediation of message credibility[J]. Library & information, 2020(5): 39-47.)
[17] 钱旦敏, 郑建明, 王文敬, 等. 基于信息可信度评估的突发公共卫生事件谣言识别研究[J]. 情报科学, 2024, 42(2): 35-42. (QIAN D M, ZHENG J M, WANG W J, et al. Rumor identification of public health emergency based on information credibility assessment[J]. Information science, 2024, 42(2): 35-42.)
[18] LIU X, ZHANG L, SUN L, et al. Survival analysis of the duration of rumors during the COVID-19 pandemic[J]. BMC public health, 2024, 24(1): 519.
[19] KAUFHOLD M A, BAYER M, HARTUNG D. et al. Design and evaluation of deep learning models for real-time credibility assessment in twitter[J]. Lecture notes in artificial intelligence, 2021, 12895: 396-408.
[20] AHMAD F, RIZVI S. Identification of credibility content measures for twitter and Sina-Weibo social networks[C]// Proceedings of ICETIT 2019: Emerging trends in information technology. Berlin: Springer, 2020, 605:372-384.
[21] 李一铭, 徐绪堪, 王普查. 面向突发事件的网络舆情可信度评估及政务微博引导研究[J]. 情报杂志, 2021, 40(11): 87- 92, 120. (LI Y M, XU X K, WANG P C. Research on credibility evaluation of network public opinion and government microblog guidance for emergencies[J]. Journal of intelligence, 2021, 40(11): 87-92, 120.)
[22] 王梓宁, 张国防. 考虑多交互关系与情感倾向的微博用户可信度评估算法[J]. 计算机应用研究, 2024, 41(10): 3000- 3007. (WANG Z N, ZHANG G F. Microblog users’ credibility evaluation algorithm considering multi-interaction relationships and emotional tendencies[J]. Application research of computers, 2024, 41(10): 3000-3007.)
[23] HOVLAND C I, WEISS W. The influence of source credibility on communication[J]. Public opinion quarterly, 1951(15): 635-650.
[24] 朱梦珺, 蒋洪迅, 许伟. 基于金融微博情感与传播效果的股票价格预测[J]. 山东大学学报(理学版), 2016, 51(11): 13-25. (ZHU M J, JIANG H X, XU W. Weibo moods and propagation factors based stock prices prediction[J]. Journal of Shandong University(natural science), 2016, 51(11): 13-25.)
[25] 陆泉, 岳雪琪, 刘婷, 等. 基于多维特征的网络健康信息可信度评估研究[J]. 信息资源管理学报, 2021, 11(3): 121-131. (LU Q, YUE XQ, LIU T, et al. Research on the credibility assessment of network health information based on multidimensional information characteristics [J]. Journal of information resource management, 2021, 11(3): 121-131.)
[26] PAUL V S, KAREN R, CHRISTOPHER W, et al. Risk as effect: the effect heuristic in cybersecurity [J]. Computers & security, 2020(90): 101651-101651.
[27] 曹碧薇, 曹玖新, 桂杰, 等. 面向中文文学作品的长文本人物关系抽取[J]. 中文信息学报, 2023, 37(5): 88-100. (CAO B W, CAO J X, GUI J, et al. Character relation extraction from Chinese literature[J]. Journal of Chinese information processing, 2023, 37(5): 88-100.)
[28] 袁平宇, 邱林. 基于BERT的多模型融合的Web攻击检测方法 [J]. 计算机工程, 2024, 50(11): 197-206. (YUAN P Y, QIU L. Web attacks detection method based on BERT with multi-model fusion[J]. Computer engineering, 2024, 50(11): 197-206.)
[29] 王振琦, 陈涛, 张宝宇, 等. 基于交叉特征感知融合的多领域虚假新闻检测[J]. 计算机系统应用, 2024, 33(3): 264- 272. (WANG Z Q, CHEN T, ZHANG B Y, et al. Multi-domain fake news detection based on cross-feature perception fusion[J]. Computer systems & applications, 2024, 33(3): 264-272.)
[30] 冯兰萍, 董陈超, 徐绪堪. 基于混合神经网络的突发公共卫生事件微博谣言识别研究[J]. 情报杂志, 2022, 41(12): 81- 88. (FENG L P, DONG C C, XU X K. Research on Weibo rumor identification in public health emergencies based on hybrid neural network[J]. Journal of intelligence, 2022, 41(12): 81-88.)
[31] 黄学坚, 王根生, 罗远胜, 等. 融合多元用户特征和内容特征的微博谣言实时检测模型[J]. 小型微型计算机系统, 2022, 43(12): 2518-2527. (HUANG X J, WANG G S, LUO Y S, et al. Weibo rumors real-time detection model based on fusion of multi user features and content features[J]. Journal of Chinese computer systems, 2022, 43(12): 2518-2527.)
Options
文章导航

/

〈 〉