基于动态数据的灰色关联分析法构建中医体质动态变化规律知识图谱

The constitution of traditional Chinese medicine (TCM) is affected by many factors such as congenital heredity, social environment and psychological character. Previous researches of this project and the similar studies at home and abroad have found that there is a correlation between individual constitutional type and some factors at some point, but there is still a lack of dynamic variation analysis of individual constitution and the key influencing factors of individual constitution. Grey relational analysis is a method to analyze the degree of influence to which factors affect the subjects, and knowledge graph is a method to visualize knowledge semantic network based on mathematical algorithm. On the basis of previous research, the hypothesis that "based on the dynamic data of physique, the key influencing factors of physique can be revealed and the dynamic changing rules of physique can be explored by using the grey correlation analysis method and knowledge map technology" is established. .On the basis of the TCM constitution nine points method, with scale identification method as the main method and traditional four-diagnosis identification method as the auxiliary method, this project proposes to collect 19200 TCM constitution information from southwest using the mobile Chinese medicine constitution identification system developed in the previous researches in a dynamic and Objective way, and to form the 12000 valid TCM constitution information through the data preprocessing. It’s expected to reveal the key influence factors of TCM constitution types, and to explore the dynamic variation rule of TCM constitution by using the knowledge graph technology and the gray relative analysis method. It will provide the quantitative analysis and the scientific proof for guidance to treat disease based on TCM constitution, and guide the preventive treatment of disease in clinical in a more scientific and reasonable way.

中医体质受先天遗传、社会环境、心理性格等多因素影响。课题前期和国内外同类研究发现某一时刻下个体体质类型与某些因素具有相关性，但尚缺乏对个体体质的关键影响因子及个体体质的动态变化分析。灰色关联分析法是一种分析因子对主体影响程度的方法，知识图谱是一种基于数学算法来可视化知识语义网络的方法。结合前期研究基础，由此建立了“基于体质动态数据，运用灰色关联分析法和知识图谱技术，揭示体质关键影响因子，探索体质动态变化规律”的假说。.本课题拟依据中医体质九分型，在前期研发的移动中医体质辨识仪的基础上，以量表辨识法为主、传统四诊辨识法为辅，动态、客观采集西南地区19200份中医体质辨识信息，通过对数据的预处理形成12000份有效样本，运用知识图谱技术和灰色关联分析法，以期揭示中医体质类型的关键影响因子，探究中医体质动态变化规律，为基于中医体质指导治未病提供量化分析和科学依据，以更科学、合理地指导临床治未病。

中医“体质辨识是“治未病”的抓手，通过辨识体质类型，以有的放矢地防病养生。体质对疾病的发生和转变起重要作用，体质受先天遗传、后天年龄、性别、气候环境、生活起居、饮食习惯等多因素影响，影响因素的改变会引起个体体质的变化。项目前期和国内外同类研究发现某一时刻下个体体质类型与某些因素具有相关性，但尚缺乏对个体体质的关键影响因子及个体体质的动态变化分析。. 项目按量表辨识法和人工辨识法约8：2的比例动态采集个体24个节气中医体质辨识信息，具体过程如下：拟定伦理审查相关资料、申请伦理审查、与每位受检者签订《知情同意书》、动态采集量表辨识法体质辨识信息27168份，人工体质辨识信息3523份。对采集数据经过预处理、知识库构建、灰色关联法分析、知识图谱构建、专家调查等研究路径，形成了中医体质因果、兼夹、传变三种关系知识图谱，揭示了中医九大体质关键核心影响因子。通过分析量表法9大体质影响因子灰色关联度分析发现，宗教信仰、民族、出差频次、家庭氛围、居住环境、过敏史、出生方式均为9大体质的前7大关键影响因子。通过分析人工体质辨识法的9大体质灰色关联度分析发现，躯体正常度、五官正常度、现病史、家庭氛围等为体质类型的关键影响因素。通过分析体质类型间的传变发现，8大偏颇体质通过体质调摄均可转为平和质，偏颇体质间亦存在一定的传变关系。通过分析兼夹情况发现，基本平和质，有血瘀质、湿热之、痰湿质、阳虚质、气虚质、气郁质倾向；8大偏颇体质间亦存在部分兼夹情况。. 本项目的研究阐释了中医体质动态变化规律，有助于更好地指导防病养生，对中医治未病具有重要意义，亦可弥补当前研究的不足，为中医体质动态变化规律等关键问题研究提供新的思路和方法学参考。