王东丽, 邵帅, 丁磊, 陆玉梅, 丁菊红, 严颖. DeepSeek语言模型对护理个案理解能力测试与应用展望. 2025. biomedRxiv.202512.00048
DeepSeek语言模型对护理个案理解能力测试与应用展望
通讯作者: 邵帅, 1508220133@qq.com
DOI:10.12201/bmr.202512.00048
Evaluation of the DeepSeek Language Model’s Comprehension of
Corresponding author: shaoshuai, 1508220133@qq.com
-
摘要:目的:探讨DeepSeek对护理学知识的理解和应用能力,评估其在护理领域中的应用前景。方法:从《护理综合案例分析》中选取护理个案,共计120个问题作为测试内容。使用DeepSeek生成问题答案,再由两名研究人员分别对模型答案和答案解析进行独立评分。通过对比DeepSeek与答案解析在准确性、全面性、实用性、简洁性、条理性及总分维度的得分情况,评估其对护理学知识的理解与应用能力。结果:DeepSeek与答案解析的得分均较高,且DeepSeek在全面性、条理性和总分的得分高于答案解析,差异具有统计学意义(p<0.05)。在准确性、实用性和简洁性方面,DeepSeek的得分低于答案解析,差异具有统计学意义(p<0.05)。结论:DeepSeek能够准确理解并回答护理学相关问题,表现出较强的护理知识理解能力。其在护理领域具有较大的应用潜力,能够有效提高护理人员的工作效率,推动护理服务向智能化方向发展。
Abstract: Objective:To investigate the DeepSeek language model’s ability to comprehend and apply nursing knowledge, and to evaluate its potential applications in the field of nursing.Methods:A total of 120 questions were selected from Comprehensive Nursing Case Analysis as test material. The DeepSeek model was used to generate answers to the questions, and two researchers independently scored both the model-generated responses and the standard answer explanations. The evaluation focused on six dimensions: accuracy, comprehensiveness, practicality, conciseness, organization, and overall score, to assess the model’s understanding and application of nursing knowledge.Results:Both DeepSeek and the standard explanations achieved high scores. DeepSeek scored significantly higher than the standard answers in terms of comprehensiveness, organization, and total score (p < 0.05). However, in terms of accuracy, practicality, and conciseness, DeepSeek scored significantly lower (p < 0.05).Conclusion:DeepSeek demonstrates a strong ability to understand and respond to nursing-related questions, indicating a high level of comprehension of nursing knowledge. It shows promising application potential in the nursing field, with the capacity to improve nurses’ work efficiency and promote the intelligent transformation of nursing services.
Key words: DeepSeek; Artificial Intelligence; Language Model; Nursing Cases; Nursing Knowledge; Evaluation Study; Model Assessment提交时间:2025-12-17
版权声明:作者本人独立拥有该论文的版权,预印本系统仅拥有论文的永久保存权利。任何人未经允许不得重复使用。 -
图表
-
牛宇翔, 葛珊衫, 王力华. 从传统NLP到大语言模型电子病历生成技术的探索与研究. 2024. doi: 10.12201/bmr.202412.00080
武凡祺, 许超. DeepSeek赋能医学教育:特点、影响与应对. 2025. doi: 10.12201/bmr.202504.00033
杨亚洲, 郑石林, 周小毅, 章施静, 赵坚. 基于知识图谱的HIV防治大模型RAG问答系统构建研究. 2025. doi: 10.12201/bmr.202511.00061
谢甲琦, 李政. 基于预训练语言模型的公众健康问句分类. 2021. doi: 10.12201/bmr.202101.00017
薛惠元, 张永高. 共同富裕视域下我国长期护理保险政策评价与优化路径研究——基于PMC指数模型. 2023. doi: 10.12201/bmr.202306.00018
马傲, 葛小玲. 人工智能大模型在医疗健康领域应用的研究. 2024. doi: 10.12201/bmr.202408.00039
康一帅, 邵陈杰. 基于统一语言模型的中医文献问题生成算法. 2022. doi: 10.12201/bmr.202110.00044
向艾宁堃, 田靖雪, 胡德华, 刘海霞. 生成式人工智能对于老年糖尿病应答效能的比较研究. 2025. doi: 10.12201/bmr.202503.00019
汪亚男, 徐霞艳, 朱月兰. “互联网+护理服务”质量管理评价指标研究的范围综述. 2024. doi: 10.12201/bmr.202407.00067
李日丽, 潘嘉明, 容世强, 孙晓翠, 易法令. 知识图谱与大语言模型融合在糖尿病健康管理中的应用进展. 2025. doi: 10.12201/bmr.202508.00035
-
-
公开评论 匿名评论 仅发给作者
引用格式
访问统计
- 阅读量:42
- 下载量: 0
- 评论数:0

登录
注册




京公网安备