网络辨证的算法原理

Algorithms and Fundamental of Network Differentiation

中医学从系统论的角度整体认识生命与疾病,钱学森多次强调“人体科学一定要有系统观,而这就是中医的观点”。课题组将中医学的病机概念与图论的有向图概念进行整合,提出了病机网络的概念,并在系统科学指导下,结合图论理论提出一种以病机为核心的数字化辨证新方法——网络辨证。本文将从数学层面详细阐述网络辨证的算法原理。网络辨证医理模型包括三个阶段,第一阶段基于笛卡尔积操作穷尽枚举所有的候选病机网络,第二阶段基于中医的整体观原则筛选出有限的整体性网络;第三阶段基于人机交互策略确定唯一的个体化病机网络。研究过程中我们基于Dijkstra算法对中医的整体观思想建立了启发式的数学化定义,并提出且证明了三个重要的定理推论对该定义进行算法层面的优化。总结而言,网络辨证与中医实践者的辨证思维是高度吻合的,网络辨证为中医数字化辨证提供了新思路。

Traditional Chinese medicine(TCM)researches the body as a whole which has the immanent connection with systems thinking.One of the most famous Chinese scientists Dr.Qian Xuesen had put forward that“the human body system is an open complex system and TCM is a complex science”.By integrating the concept of pathogenesis in the terminology of traditional Chinese medicine(TCM)with the concept of directed graph in the terminology of graph theory,a novel concept,pathogenesis network,was proposed in our previous study.Furthermore,a computational method of pathogenesis inference,network-based differentiation,is proposed by utilizing the graph-theory.Our method is a three-stage approach.The first-stage is to enumerate all the possible pathogenesis network via Cartesian production,the second-stage is to screen some validated pathogenesis networks via a heuristic approach,and the third-stage is to pick out the personalized pathogenesis network via practitioner-computer interaction.In the second stage,a philosophical concept which pervades ubiquitously in the TCM theory,holistic thinking,is mathematically formulated by Dijkstra algorithm to scree the validated networks from all the possible networks.Besides,we state and prove some important theorems for the optimization of our heuristic approach in the second stage.In general,our computational method,network-based differentiation,is highly consistent with the practitioners’perception,which suggests ideas for the intelligent differentiation of pathogenesis in TCM.