Research On Visual Identification Method Of Health State Pattern In Traditional Chinese Medicine Based On Granules Partial Order Structure

Traditional Chinese medicine(TCM)has left abundant medical and information resources of diagnosis and treatment of diseases and health care for mankind.Promoting the development of TCM has become a national strategy.Using modern artificial intelligence technology and knowledge discovery methods to exploit the knowledge contained in TCM medical records is of great significance to the development of TCM.However,in the process of inheriting TCM culture and modernization of TCM,there are several problems to be solved: TCM theories are mostly described in natural language,while modern science and technology is mostly described in formalized mathematical language,which makes a natural barrier for the integration of TCM and modern science and technology;The highly subjective process of TCM syndrome differentiation and treatment and a certain degree of fuzziness and imprecision in TCM data generate a urgent requirement for corresponding knowledge discovery methods to complete the tasks of TCM data processing and TCM knowledge discovery;TCM data contains rich implicit information,the research on effective measures of displaying the relationship between TCM concepts including the explicitation of tacit knowledge can facilitate and accelerate the inheritance and modernization of TCM knowledge.Facing the related problems in the process of modernization of TCM,this dissertation conducts the study on the aspects of formal description of TCM theory,visual knowledge discovery based on granular partial order formal structure,and intelligent identification of TCM human health state pattern and knowledge visualization of TCM syndrome element pattern.The theory and method of intelligent identification of human health state in TCM based on the generation principle of granular partial order formal structure are preliminarily constructed,and verified by using the actual medical case data of TCM.The research contents of this dissertation are as follows:Firstly,based on the disciplinary essence of "phenomenon-state medicine" of TCM,the theory of TCM diagnostics and human health state are represented by abstract conceptual model and formal mathematical model,and the formal language description of TCM diagnostic thinking is preliminarily completed.This dissertation illustrates the mathematical principle of "inter domain mapping and sub domain coverage" in TCM human health state identification,and proposes a computable model for TCM human health state identification based on the construction of human complex system model and syndrome element differentiation theory.Secondly,from the perspective of granular computing,the research proposes a knowledge discovery method based on the visual representation principle of granular partial order formal structure.It presents a unified representation architecture of heterogeneous data based on data granulation and formal information table,and puts forward the methods of granular partial order formal structure generation,visual representation and knowledge discovery.Under the condition of introducing the data pattern similarity measurement method based on Boolean attribute distance,a knowledge discovery method architecture based on granular partial order formal structure visual representation and Boolean attribute distance is proposed,and its effectiveness is verified by experiments.Thirdly,the knowledge discovery method based on granular partial order formal structure visual representation and Boolean attribute distance is extended to the field of semi-supervised learning,and a self-training semi-supervised learning method based on granular partial order formal structure and Boolean attribute distance is proposed,which provides a new idea for knowledge discovery in TCM and other fields under the background of less labels.Finally,on the basis of TCM medical case data based on the principle of inquiry,using the proposed knowledge discovery method based on granular partial order formal structure,an intelligent identification method of TCM human health state is established,so as to realize the construction of computable model of TCM human health state identification and fulfill the knowledge discovery of TCM human health status and TCM syndrome element pattern distribution rule of the sample population.