Methodological Research On The Construction Of Evaluation Index System Of Traditional Chinese Medicine Medical Case Effect Based On Static Bayesian Network

Traditional Chinese Medicine medical records are records of medical doctors'clinical experience,recording the doctor's description,diagnosis,treatment plans and treatment results of clinical problems,reflecting and verifying medical opinions of doctors,inspiring new theories and guiding clinical practice effect.In the process of clinical diagnosis and treatment.in the face of a patient's specific symptoms,experienced doctors can often combine past medical records to assist in solving the current clinical problems-if the clinical problems recorded in a medical record are similar to the current patient's clinical problems,Perhaps the diagnosis and treatment ideas or treatment plans can be reused.Among them,the effectiveness of the previous treatment plan is a key factor in whether the current clinical problem can be effectively solved.The effectiveness evaluation of a Traditional Chinese Medicine medical record treatment plan is an evaluation of the effect of the treatment plan under the actual medical conditions in the past,The evaluation result includes both the effectiveness of the treatment plan and the influence of other external factors.It belongs to the category of effect evaluation.At this stage,the research methods of Traditional Chinese Medicine medical records mainly include comprehensive analysis,statistical analysis,and Case-based Reasoning.Case-based Reasoning has been introduced into the field of Traditional Chinese Medicine because of its"reasoning method is consistent with the thinking process of Traditional Chinese Medicine syndrome differentiation and treatment" and has become Traditional Chinese Medicine medical records.However,at this stage,whether it is traditional medical case research methods or Casebased Reasoning,the standardization and process-based processing of the effect evaluation of Traditional Chinese Medicine medical records have not been realized.This is related to the lack of a unified and standardized method of constructing an effect evaluation index system for Traditional Chinese Medicine medical records.Through exploring the construction method of Traditional Chinese Medicine medical record effect evaluation index system based on static Bayesian network and carrying out case studies,it hopes to provide methodological reference for the construction of Traditional Chinese Medicine medical record effect evaluation index system.1.Research purposeWith the goal of assisting clinical decision-making,analyze the key link in the evaluation of the effect of Traditional Chinese Medicine medical records—difficulties in the construction of the effect evaluation index system,and build a static Bayesian network-based evaluation index system for the effect of medical records,hoping to be the evaluation index of the effect of medical records System construction provides methodological reference.2.Research methodsStatic Bayesian network is a graphical description of probabilistic relationships,which is mainly composed of two parts:network structure and parameters.The network structure reflects the qualitative characteristics of domain knowledge(represented by node variables and the association relationship between variables).Network parameters reflect the quantitative characteristics of domain knowledge(quantified by the conditional probability distribution between node variables).This research uses medical records as the original data,and mainly obtains the evaluation index system of medical records through two links:(1)Use static Bayesian network structure learning tools to develop Bayesian network structure learning,and obtain qualitative evaluations between evaluation indicators.Association relationship;(2)On the basis of structural learning,carry out static Bayesian network parameter learning to obtain a quantitative description of the dependence relationship between evaluation indicators.3.Research results(1)Methodological level:Guided by the construction method of static Bayesian networks,combined with the characteristics of medical records,a relatively complete method and process of constructing a Traditional Chinese Medicine medical record effect evaluation index system based on static Bayesian networks has been formed.The methodology process includes six main links,namely:data collection and standardization,node selection,structure learning method selection,parameter learning method selection,Bayesian network learning tool selection,reliability analysis and validity test.(2)Case study:Take Diabetes(type 2 diabetes)as an example,construct a Diabetes Diabetes evaluation index system based on a static Bayesian network:?The evaluation index system contains a total of 18 indicators,namely," urine color,Stool texture,fatigue,...,fasting blood sugar,fur quality(moisture dryness),etc.";?There are certain conditional dependencies between evaluation indicators,such as "thirst" conditions depend on "tongue shape(fat and thin)" and "urination" The conditions of "color" and "fasting blood glucose" depend on "moss quality(thickness)" and "moss quality(moisture)";?The dependence relationship between indicators is quantitatively expressed in the form of conditional probability,such as when the coating is thick and moist.When the fasting blood glucose was taken as "normal,+,++,+++",the probabilities were 0.02,0.06,0.48,and 0.44,respectively.That is,the construction of the evaluation index system for the effect of diabetic medical records based on the static Bayesian network can qualitatively describe and quantitatively express the dependence between the evaluation indexes.4.Research conclusionThe research verifies the methodology process through the collection and processing of medical records of Diabetes(type 2 diabetes),static Bayesian network structure learning and parameter learning,and model verification.Achieved the construction of an evaluation index system for the effect of diabetic disease based on the static Bayesian network.lt is believed that the medical record effect evaluation index system based on the static Bayesian network can intuitively and clearly describe the dependence relationship between the evaluation indicators and the size of the dependence relationship.This presentation method is more in line with the concept and clinical practice of Traditional Chinese Medicine syndrome differentiation and treatment,and it also provides a direction for further exploring the relationship between symptoms,signs,and physical and chemical indicators,and rationally using these indicators to participate in the effect evaluation.The amount of data will have an impact on the accuracy of the model results,but it can be compensated by increasing prior knowledge and selecting the corresponding "small sample" learning algorithm.Moreover,the static Bayesian network will be continuously updated with the increase of the amount of data and tends to a stable value,which is also one of the advantages of using the static Bayesian network to construct the effect evaluation index.The evaluation index system construction method based on the static Bayesian network is not only suitable for the construction of the effect evaluation index system of Traditional Chinese Medicine medical records,but also can provide ideas and methodological references for the construction of the effect evaluation index system in the clinical diagnosis and treatment of Traditional Chinese Medicine.