中医智能诊疗研究中的数据标准化瓶颈和处理策略

Data standardization bottlenecks and processing strategies in traditional Chinese medicine intelligent diagnosis and treatment research

中医辨证论治过程是一个以医师为主体、医患互动的,动态和个体化的思维和实践过程,主观因素在其中具有重要作用,这是数据标准化问题的主要原因,多种辨证体系和学术流派并存加剧了这一问题,具体表现为数据元名义不统一且难以量化、结构化程度低且多源异构。中医诊疗数据的处理策略,根据处理对象类型的不同,可分为复合义项拆分、缺失项补充、同名异义项区分、异名同义项统一等;根据标准化处理与数据采集先后不同可分为前标准化、同步标准化和后标准化处理等。中医药诊疗数据标准日趋完善,标准兼容性和可应用性日益提高,标准化预处理的智能化程度不断加强,为数据标准化的自动化、后台化和高效化提供可能,在数据标准化基础上的个性化和精准化,将成为中医诊疗疗效稳定和诊疗过程可重复的保障。

The process of syndrome differentiation and treatment in traditional Chinese medicine(TCM)is a dynamic and individualized thinking and practice process with physicians as the main body and doctor-patient interaction,in which subjective factors play an important role,which is the main reason for the problem of data standardization,and the coexistence of multiple syndrome differentiation systems and academic schools exacerbates this problem,which is manifested in the fact that the name of the data element is not unified and difficult to quantify,the degree of structure is low and multi-source heterogeneous.Processing strategy of TCM diagnosis and treatment data can be divided into compound semantic terms splitting,missing items supplemented,homonymous and synonymous items distinguished,and different names and synonyms unified.According to the different sequences of standardized processing and data collection,it can be divided into pre-standardization,synchronous standardization,and post-standardization processing.The data standards of TCM diagnosis and treatment are becoming more and more perfect,the compatibility and applicability of the standards are improving day by day,and the intelligence of standardized preprocessing is continuously strengthened,which provides the possibility of automation,background and efficiency of data standardization,and the personalization and precision on the basis of data standardization will become the guarantee of the stability of TCM diagnosis and treatment efficacy and the reproducibility of the diagnosis and treatmentprocess,