The Data Mining And Constructing A Syndrome Differentiation Model Of Coronary Heart Disease Cases Treating By Distinguished Veteran TCM Doctors

Objective: Analyzing the coronary heart disease cases treating by distinguished veteran TCM doctors with statistical methods and data mining technology.To exploring the therapy of distinguished veteran TCM doctors curing coronary heart disease and providing bases for treating coronary heart disease by Chinese medicine. Constructing a TCM syndrome differentiation model of coronary heart disease and providing assistance for the syndrome differentiation of coronary heart disease.Methods:626 cases of coronary heart disease treated with distinguished veteran TCM doctor in 1980 s were recorded in the case analysis system. Analyzing syndrome differentiation rule, lingua and pulse rule and pharmacy rule applying the case analysis system and SAS statistical software, doing data mining applying Chinese medicine prescription analysissystem. Constructing a syndrome differentiation model of coronary heart disease applying BP neural network and Matlab software.Results:(1)Frequency Analysis:①Type of Syndrome: The common high frequency type of syndrome are blood stasis, deficiency of qi,deficiency of yang,phlegm and retained fluid,stagnation of the circulation of qi,deficiency of yin,retained fluid and theyoften appear coherently; ② Symptom: The thoracic oppression,pectoralgia, palpitation, fatigue, dizziness and wheeze are more;③Tongue Nature: The light tongue,dim tongue, swollen tongue, teeth-marked tongue,ecchymosis tongue and red tongue are more;④Coated Tongue: The thin coating and white coating are more;⑤Pulse Condition: The deep pulse, taut pulse and thready pulse are more.(2)Canonical Correlation Analysis:①Symptom:The deficiency of qi and fatiguehave typical interrelated relation;②Tongue Nature:The blood stasis and ecchymosis tongue, deficiency of blood and light tongue, phlegm and retained fluid and swollen tongue, deficiency of yin and red tongue have typical interrelated relation;③Coated Tongue:The phlegm and retained fluid and white coating, muggy and yellow coating have typical interrelated relation;④Pulse Condition: The retained fluid and slippery pulse, stagnation of the circulation of qi and taut pulse have typical interrelated relation.(3)Pharmacy Analysis:①Pharmacy Frequency:There are 261 traditional Chinese medicine in the 626 cases, in which salvia miltiorrhiza is the most one and semen sojae praeparatum, terra flava usta and someother traditional Chinese medicine are least ones;②Cluster analysis:There are 12 kinds of 45 traditional Chinese medicine which have appled more than 40;③Discriminate Pharmacy: Analyzing pharmacy of syndrome differentiation for blood stasis, deficiency of qi,deficiency of yang,phlegm and retained fluid,stagnation of the circulation of qi,deficiency of yin and retained fluid of coronary heart disease cases by logistic regression analysis,then get the Logistic regression equation; ④ Symptom Pharmacy: Analyzing pharmacy of different symptom for thoracic oppression,pectoralgia, palpitation, fatigue, dizziness, wheeze,lassitude, edema, chills,distention and fullness, constipation, anorexia, headache, thirst, insomnia, upset, abdominal fullness, spontaneous sweating, hypochondriac pain and purple lips of coronary heart disease cases by logistic regression analysis,then get the Logistic regression equation;⑤Data mining: Using the improved Apriori algorithm, and according to the least confidence level is 0.1, data mining for all the coronary heart disease cases, coronary heart disease cases of 7 kinds of common types of syndrome, coronary heart disease cases of 8 kinds of common clinical symptoms respectively, get different number of 1-25 prescriptions respectively.(4)The construction of the syndrome differentiation model: The forecast accuracy of each syndrome differentiation model are blood stasis 90%, deficiency of qi 95%.Conclusion: With the help of statistical and data mining technology including canonical correlation analysis, cluster analysis, logistic regression analysis, improved Apriori algorithm, this study analyzed the cases of coronary heart disease treated with distinguished veteran TCM doctor systematically and objectively, the results have clinic significance and can apply impersonal bases for treating coronary heart disease of traditional Chinese medicine. Constructed syndrome differentiation model of coronary heart disease applying BP neural network, which could provide assistance for the syndrome differentiation of coronary heart disease, at the same time, they can apply new method for inheriting experience of distinguished veteran TCM doctors.