现代情报2025，Vol.45Issue(10)：115-124,10.DOI:10.3969/j.issn.1008-0821.2025.10.010
智能诊断模型的可解释性策略研究
Research on Interpretability Strategy of Intelligent Diagnostic Model
摘要
Abstract
[Purpose/Significance]In the context of disease diagnosis,this paper explores interpretability strategies for AI diagnostic models based on the integration of knowledge and data,in order to promote the development of intelligent medicine and further strengthen the cross integration between information science,data science,and medical disciplines.[Method/Process]Combining theories and methods from multiple disciplines,the study constructed a theoretical frame-work for interpretable intelligent diagnosis models from the perspective of dual-wheel driving by both knowledge and data,and conducted an empirical study on the differential diagnosis of viral and bacterial pneumonia.[Result/Conclusion]From the perspective of informatics,this paper elaborates on the construction strategies of interpretable AI diagnostic models and verifies the feasibility and rationality of the framework through practical application scenarios.关键词
双轮驱动/人工智能/可解释性/智能诊断/肺炎Key words
dual-driven/artificial intelligence/explainability/intelligent decision-making/pneumonia分类
医药卫生引用本文复制引用
于海涛,牟冬梅,王长聪..智能诊断模型的可解释性策略研究[J].现代情报,2025,45(10):115-124,10.基金项目
国家自然科学基金项目"'知识—数据'双轮驱动的多模态电子病历数据组织模式与应用研究"(项目编号:72374081) （项目编号:72374081）
吉林省科技发展计划项目"面向健康自主管理的个人诊疗信息平台开发与应用"(项目编号:20240304164SF). （项目编号:20240304164SF）
