基于有向图的证素归因模型构建ISO-R证素辨证体系的方法学研究—以肺系疾病辨证为例

The digraph is an important nonlinear data structure , and TCM has obvious nonlinear characteristics. (Classic)Syndrome factor differentiation is an important method system of TCM intelligent syndrome differentiation in recent years.Based on this, the research group in the prophase project has integrated the concept of digraph and syndrome factor, established the concept of syndrome factor-digraph ,proposed the syndrome factor combination of attribution of formation pathogenesis of universal law -—ISO - R law. This topic will further improve the relevant concepts and contents of the syndrome factor-digraph. Through the study of the relationship between symptom and syndrome factor status (i.e., there is relationship or no relationship), we avoid the numerical relationship between symptom and syndrome weights in classical syndrome differentiation system. Through the dynamic construction of syndrome factor combination and objective screening of the ISO-R law to establish a new system of three - order double - net syndrome factor differentiation called as"symptom-syndrome factor digraph-pathogenesis", and expresses the pathogenesis in the form of syndrome factor-digraph to avoid the explosion of static syndrome factor produced by the standard syndrome pattern in the classical syndrome differentiation system.finally to construct a TCM intelligent syndrome differentiation model with symptoms set as input and pathogenesis (shown as syndrome factor-digraph)as output ;The research group will take the syndrome differentiation of multiple diseases domain (lung diseases) as an example, complete the relevant knowledge base, and write the relevant dialectical procedure，dialectical practice man-machine cooperation pattern implementation through real cases, and the results with the traditional human brain syndrome, repeatedly train and optimize the new syndrome differentiation system, ultimately for traditional Chinese medicine with mathematics and computer crossover provides new idea.

有向图是重要的非线性数据结构，中医学具有明显的非线性特性；（经典）证素辨证是近年来探索中医智能辨证的重要工具体系。基于此，课题组在前期研究中将有向图与证素的概念进行整合，建立了证素有向图的概念，提出了证素组合“归因串联”形成病机的标准—ISO-R法则。本次课题拟进一步完善证素有向图的概念，通过症状与证素的状态关系研究，回避经典证素辨证体系中症状与证素计量权值的数值关系研究；通过证素组合的动态构造与ISO-R法则的动态筛选，建立“症状—证素有向图—病机”的三阶双网证素辨证体系，并采用证素有向图的形式动态表达病机，回避经典证素辨证体系中采用标准证名模式表达病机而产生的静态证素组合爆炸问题。最后以肺系疾病为例，构建知识库，编写辨证程序，通过人机协作模式对真实病例进行辨证实践，并将该结果与传统人脑辨证结果对照，反复训练与优化该辨证新体系，最终为中医智能辨证提供新思路。

中医临床病机具有深刻的复杂性、传变性，常涉及多个病位、病因、病性等病机要素,且病机要素之间相互联系、互为因果,具有典型的系统性属性。受项目资助，课题团队以系统科学为指导并结合图论理论,首次提出一种辨病机的数字化辨证方法—网络辨证。本研究的创新点如下：（1）借助图论有向图结构，首次提出病机的网络化定义—病机网络。（2）明确病机网络的整体性定义，将中医的整体观(念)数学化为整体性(质)；（3）从工程技术的角度实现网络辨证算法原理。. 网络辨证医理模型包括三个阶段，第一阶段基于S2P(症状-病机元关系)的笛卡尔积运算并通过P2P(病机元-病机元关系)的关系映射生成所有的候选病机网络,第二阶段基于网络整体性定义筛选出有限的整体性网络；第三阶段基于人机协作策略确定出唯一的临床病机网络。. 课题组以单病域（肺系疾病）为研究范式，开展了网络辨证的实证应用，并通过100例真实临床病例开展仿真辨证，研究过程中，项目组提出了一致性评价策略，仿真辨证结果显示基于5组人机协作的病机网络具有高度一致性(病机元一致性为89.81%,病机链一致性为91.22%)，网络辨证能够充分拟合中医实践者的推理思维。. 《中医杂志》对网络辨证撰有如下评述：信息技术在中医药现代化进程中正在并将继续发挥重要作用。网络辨证是以系统科学为指导，结合图论理论而建立的一种以病机为核心的智能辨证新方法。该辨证方法突破了中医辨证型的诊疗模式，强调中医辨病机的核心思维，通过对病机的网络化定义，拟合了临床病机深刻的复杂性、传变性、非线性等科学内涵，还原了中医辨证过程中固有的系统观、整体观、个体观等推理思维。网络辨证守“病机”之正，创“智能”之新，通过多学科交叉为中医数字化辨证提供了思路，是创建符合中医原创思维的全新诊疗体系的创新和实践。

内容获取失败，请点击重试