PRDAGE:基于数据增强和多图嵌入的传统中医药的处方建议框架
Zhihua Wen1, Yunchun Dong2, Lihong Peng1
1School of Computer Science, Hunan University of Technology, Zhuzhou, China.
PeerJ. Computer science
|September 24, 2025
概括
这项研究介绍了PRDAGE,这是传统中医药 (TCM) 处方建议的新框架. 通过整合语义信息和使用数据增强来提高模型训练,PRDAGE提高了准确性.
科学领域:
- * 传统中国医学 (TCM)
- * 计算医学是一种医学.
- * 数据科学数据科学
背景情况:
- * 传统中医 (TCM) 处方对于维持健康和治疗疾病至关重要.
- *现有的TCM处方建议研究主要关注症状与草药的相关性,忽视语义信息.
- *在TCM研究中,有限的数据集大小阻碍了有效的模型培训和性能.
研究的目的:
- * 为了解决当前TCM处方推方法的局限性.
- * 开发一个框架,捕获和利用症状和草药的语义信息.
- * 提高TCM处方推模型的准确性和概括能力.
主要方法:
- * 开发了PRDAGE (基于数据增强和图形嵌入的处方建议) 框架.
- * 创建了3052个标准化的经典医学病例的数据集.
- * 采用Sentence-BERT和图形卷积网络实现了多层嵌入方法.
- *采用基于中位数的随机数据增强技术来丰富数据集.
主要成果:
- * PRDAGE在未增强数据集上的基线模型相比显示出更高的性能.
- * 在Top@10.0的精度 (1.69%) 和回忆率 (3.80%) 中取得了显著的改进.
- * 除研究证实了数据增强和多层嵌入模块的积极贡献.
结论:
- *PRDAGE是TCM处方推的有效框架.
- *多层嵌入成功捕获语义信息和复杂的关系.
- *以中位数为基础的数据增强提高了模型性能和概括性.
相关概念视频
01:28Dosage Regimens: Designs and Approaches
Designing a dosage regimen, which refers to the manner of drug administration, is a complex process involving the selection of drug dose, route, and frequency. This process is underpinned by pharmacokinetic parameters derived from tests and population averages. These parameters are then tailored to patient-specific variables such as diagnosis, demographics, and allergy status. Once therapy commences, therapeutic response monitoring is critical and achieved through clinical and physical...
01:23Dosage Regimen Designs: Nomograms and Tabulations
Nomograms and tabulations are vital tools used by clinicians to design accurate and individualized dosage regimens. These instruments provide a straightforward method for adjusting dosages based on individual patient characteristics, including age, weight, and physiological condition. The foundation of a drug's nomogram is population pharmacokinetic data collected and analyzed using specific models. This data simplifies complex equations, presenting them diagrammatically or tabularly for easy...