| 高烨,李松,袁安琪.基于在线健康社区患者评论的主题识别和情感分析研究[J].情报工程,2026,(2):074-086 |
| 基于在线健康社区患者评论的主题识别和情感分析研究 |
| Research on Topic Identification and Sentiment Analysis Based on Online Health Communities Patient Comments |
| |
| DOI： |
| 中文关键词: 在线健康社区；主题识别；情感分析；医患关系 |
| 英文关键词: Online Health Communities; Topic Recognition; Sentiment Analysis; Doctor-patient Relationship |
| 基金项目: |
| 作者 | 单位 | | 高烨 | 河北大学管理学院 保定 071002 | | 李松 | 河北大学管理学院 保定 071003 | | 袁安琪 | 河北大学管理学院 保定 071004 |
|
| 摘要点击次数: 31 |
| 全文下载次数: 26 |
| 中文摘要: |
| [目的/意义]为探究在线健康社区中患者评论数据背后的医患关系失谐机理，构建医患信任修复机制，提出一种基于在线健康社区患者评论主题识别和情感分析协同研究框架。[方法/过程]首先构建逆主题频率优化LDA 模型的ITFLDA 模型用于挖掘初始主题特征词，然后引入初始主题词项间的正点互信息PPMI 重新建立特征词筛选规则，实现患者评论主题分类；最后基于主题分类结果，设计ALBERT-BiGRU 融合主题情感双重注意力机制的情感分析模型，聚焦细粒度情感判别特征，完成主题导向下的情感分类优化。[结果/结论] 通过对“39 健康网”平台患者评论数据开展实证研究，提炼出患者就诊时重点关注的主题维度：治疗效果、医术医德和诊疗态度，并深入挖掘出各主题维度下引发患者负面情绪的医患核心矛盾，据此提出针对性改进策略，为数字化医疗生态治理提供可操作性路径。 |
| 英文摘要: |
| [Objective/Significance] In order to explore the dissonance mechanism of doctor-patient relationship behind patient review data in online health communities and build a doctor-patient trust repair mechanism, a collaborative research framework based on topic recognition and sentiment analysis of patient comments in online health communities is proposed. [Methods/Processes] Firstly, the ITFLDA model that optimizes the LDA model using inverse topic frequency is constructed to mine the initial topic feature words, and then the positive point mutual information between the initial topic words is introduced to reestablish the feature word screening rules and achieved the topic classification of patient comments. Finally, based on the results of topic classification, a sentiment analysis model integrating the dual attention mechanism of topic and sentiment with ALBERTBiGRU is designed to focus on the fine-grained sentiment discrimination features, completing the optimization of sentiment classification under topic orientation. [Results/Conclusions] Through empirical research on the patient comments of “39 Health Network” platform, extract the thematic dimensions that patients focus on in the process of treatment: treatment effect, medical skill and medical ethics, and attitude towards diagnosis and treatment. The core contradictions between doctors and patients that cause negative emotions in patients under each thematic dimension are deeply explored. On this basis, targeted improvement strategies are proposed to provide an operational path for the governance of the digital medical ecosystem. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
