| The "Syndrome differentiation and treatment" process of traditional Chinese medicine is the most important part in the treatment of diseases by Chinese medicine.Traditional Chinese medicine emphasizes "treating the disease and seeking the basics".To prevent or treat a certain disease by using the theory of traditional chinese medicine,The doctor must collect the four diagnostic data of the patient and analyze them correctly.The first thing in the Syndrome differentiation and treatment is to "dialectical" accuracy.However in the process of TCM clinical syndrome differentiation,most experts use the method of "catch the main disease".Since TCM syndrome differentiation does not have the same clear diagnostic criteria as Western medicine,clinicians often rely on personal experience.Hence,the symptom capture is subjective which leads to inaccurate syndrome differentiation.At present,it is becoming a rising issue that how to help clinicians to grasp the "main disease"scientifically and effectively.1.Research purposesGuided by the theory of complexity science,this study focuses on TCM clinical symptom screening and weight labeling,taking Type 2 Diabetes Mellitus(T2DM)as an example,the author will use Decision Tree,Gibbs sampling(Gibbs Sampling)and expert opinion amendments to carry out multiple data analysis to explore the method of labeling and weight of clinical symptoms of TCM.Furthermore,screen and weight the three main symptoms of type 2 diabetes to develop a method,which can be applied in clinical treatment,to assist the decision-making process for the clinicians.2.Research methods2.1 This study used type 2 diabetes as the research object,collecting three typical clinical cases of type 2 diabetes,and establishing a separate data set for each patient.Standardize the expression of symptom terms in clinical cases according to national standardization norms and authoritative references.Based on the research methods and requirements,consolidate the clinical symptoms and build TCM symptoms database of type 2 diabetes,providing data for the study and a database of TCM symptoms of type 2 diabetes was established to provide data support for the study.2.2 Using the decision tree algorithm to screen and extract the case symptoms in the database and rank them by importance;apply Gibbs sample to analyze the results and separate and label the results.Correct the result dynamically by machine preference and expert opinions.Finally it comes out the final result with its symptoms and weight.2.3 Establish TCM symptom screening and weight labeling methods for type 2 diabetes.3.Research resultsThe entire operation process of the decision tree screened 82 symptoms,and a total of 16 models were selected to identify the symptoms that are significant for distinguishing between Qi-yin depletion,blood stasis and intestinal damp-heat.The results of the operation indicate that the tongue is dusky tongue,dry stool,thin stool,slippery pulse,dry mouth,thirst,thin coating,red tongue,acid gurgitation and heartburn,cold limbs,hearing loss,fatigue,edema and wiry pulse can play a greater guiding role in distinguishing syndrome types.According to the decision tree rules we generated,we input 91 cases of intestinal damp-heat syndrome cases.According to this rule,79 cases were accurately classified into the intestinal damp-heat syndrome group,and 9 cases were classified into Qi-yin depletion.In the syndrome group,3 cases were classified into the blood stasis syndrome group,the accuracy rate was 86.8132%;the input of 78 cases of Qi-yin depletion syndrome cases,60 cases were accurately classified into Qi-yin depletion after running according to this rule.In the deficiency syndrome group,5 cases were divided into the intestinal damp-heats syndrome group,5 cases were divided into the blood stasis syndrome group,the accuracy rate was 76.9231%;the input of 55 cases of blood stasis syndrome case data,according to this rule After 47 cases were accurately divided into the blood stasis syndrome group,5 cases were divided into the intestinal damp-heat syndrome group,and 3 cases were divided into the Qi-yin depletion syndrome group,the accuracy rate was 85.4545%.Of the 224 groups of data analyzed in this study,186 were accurately classified,with an accuracy rate of 83.0357%.There were 38 cases of unclassified case data,and the inaccuracy rate was 16.9643%.Gibbs sampling results showed that the weight of symptoms in the intestinal damp-heat syndrome was:dry mouth 7.4%,thick and greasy coating 4.8%,slippery pulse 4.5%,yellow coating 4.1%,dark urine 3.9%,thirst 3.4%,frequency urination at night 3.4%,polyuria 3.1%,red tongue 2.7%,acid gurgitation and heartburn 2.6%,polydipsia 2.5%,dizziness 2.2%,rapid pulse 2.0%,bitter taste in the mouth 2.0%,teeth-marked tongue 2.0%,halitosis 1.9%,enlarged tongue 1.9%,blurring vision 1.8%,turbid urine 1.8%,and cough 1.7%.In the syndrome of deficiency of Qi-yin depletion,the weight of symptoms was:dry mouth 5.8%,sweating 5.0%,lack of strength 4.8%,polydipsia 4.1%,polyuria 3.7%,thin coating 3.5%,thready pulse 3.2%,numbness of hand and foot and limb 3.1%,edema 3.1%,white coating 2.8%,rapid pulse 2.8%,insomnia 2.7%,limb pain 2.5%,frequency urination at night 2.4%,tongue red and lack of liquid 2.3%,turbid urine 2.3%,blurring vision 2.3%,night sweat 2.2%,weight loss 2.2%,palpitations 2.1%.The weight of symptoms in blood stasis syndrome is:dusky tongue 6.4%,numbness of hand and foot and limb 5.6%,dry mouth 5.0%,limb pain 4.5%,polydipsia 4.5%,thin coating 3.3%,wiry pulse 3.3%,blurring vision 3.1%,white coating 2.9%,frequency urination at night 2.9%,lack of strength 2.6%,stasis macules 2.5%,squamous and dry skin 2.4%,polyuria 2.4%,dizziness 2.3%,chest distress 2.3%,limb pain 2.2%,red tongue 2.1%,stroke and hemiplegia 2.0%,thirst 1.5%.Correct the results above dynamically by the machine preferences and expert opinions and obtain each main symptom and their weights.From the Gibbs sampling results,it can be figured out that among the three types of syndromes,the weights of the symptoms such as dry mouth,polydipsia,polyuria,blurring vision,and frequency urination at night are more advanced.In the Qi and Yin deficiency syndrome and blood stasis syndromeg,the numbness and weakness of the hands,feet and limbs have greater weight.According to the results,we don’t have to look at all the symptoms in a lot of complicated symptoms,such as the 0th theme derived from the Gibbs sampling process,through The typical symptoms of ranking the top weights,such as thick and greasy coating,dark urine,yellow coating,thirst,acid gurgitation and heartburn,bitter taste in the mouth,etc.,can be inferred to be intestinal damp-heat.The weighting results of several other syndrome types are also consistent with the typical symptoms of various syndrome types in the clinical diagnostic criteria,which are consistent with the clinical practice,indicating that the sampling results have practical operational significance.4.Conclusion and significanceThe TCM symptom screening and weight labeling method for type 2 diabetes designed by the study can accurately extract the symptoms that are significant for distinguishing syndrome types,and screen and mark the important symptoms under each syndrome type.This study combines of various techniques to provide a method for labeling the weight of TCM symptoms,solving the problem of weighting the main symptoms of TCM syndromes,and offering auxiliary decision-making for the clinic to help clinicians to grasp the main symptoms,and perform clinical syndrome differentiation more accurate.At the end to make the whole processes more correctly,to make the medication more reasonable,and hence benefit the final therapy. |
PDF Full Text Request