Study On "Combination Of Syndrome Factors" Of Common Diseases Of Traditional Chinese Medicine Based On Literature Mining

Objective:Through the consolidation and excavation of clinical research literatures oftraditonal chinese medicine(TCM)that published in recent 10 years,we summarized the frequency distribution of diseases,syndromes,syndrome factors in clinical research.To explore the common combination rule of TCM syndrome and provide the basis of big data analysis for the standardization research of TCM syndromes:This paper analyzes the logical decision rule based on conditional probability in the diagnosis of TCM syndrome factor,and provides the syndrome factor matching paradigm for the syndrome differentiation model in TCM health management system.Methods:1.Data sources,screening and sorting.The retrieval type of "TCM AND Clinical AND(Syndrome OR Syndrome type)"were used to retrieval China National Knowledge Infrastructure(CNKI)and VIP database(VIP)published from January 2007 to june 2016 the clinical research literature of TCM,The bibliographic information,names of diseases.syndromes and syndrome factors involved in the literature into Excel to establish a data set,and specification of terminology for the name of disease,syndrome and so on.Dismantling syndorme type to syndrome factor after standard.2.The frequency distribution of disease,syndrome type and syndrome factor which are the hot of TCM research.The frequency statistics and analysis of Excel's pivot table were used to get the frequency distribution table of common the name of chinese medicine.the name of western medicine,syndrome type,syndrome and syndrome factor.3.The common "combination of syndrome factors" in clinical research of TCM.The Apriori algorithm based on IBM Clementine 14.1 data mining platform is used to mine frequent combinations of syndrome elements,providing data support for Standardization Research of TCM syndromes.Select the hot topics of clinical research in Chinese medicine,dig out the frequent occurrence of "syndrome factor combination" rule in the entries,and provide data support for the summary of the common syndromes of hot diseases.4.Diagnosis of TCM syndrome factor.The Bayesian Network based on conditional probability were used to excavate syndrome factors diagnosis rule of TCM.to provide pattern matching thinking rule for syndrome facotr differentiation model which were achievementin Chinese medicine health management system.Results:1.9730 TCM clinical research literatures were totally included.Hot spot in clinical research of TCM according to frequency statistics from big to small,the top 10 TCM diseases ranking in sequence were chest impediment,edema,dizziness,and so on;the top 10 western medicine diseases ranking in sequence were coronary heart disease,non-insulin dependent diabetes mellitus.hypertension,and so on.The syndrome types of TCM in sequence were syndrome of dual deficiency of qi and yin,liver-kidney yin deficiency syndrome,syndrome of qi deficiency with blood stasis,and so on.The symptom factors in sequence were qi deficiency,blood stasis,heat,and so on.The syndromes of TCM in sequence were threadpulse.white fur.stringy pulse,and so on.2.The proportion of simple syndrome types was 9.40%,such as blood stasis syndrome and Qi deficiency syndrome;the cumulative proportion of syndrome types of 2-4 syndrome factors combinations was 86.04%,such as syndrome of dual deficiency of qi and yin,liver-kidney yin deficiency syndrome,syndrome of qi deficiency with blood stasis,and so on.Excavate the 24 rules of the combination rule of the syndrome element,such as liver + kidney+ Yin deficiency,spleen + kidney + Yang deficiency and so on.3.Taking common syndrome factors such as "Qi Deficiency" "Blood stasis" and "Hot",we dig out the bayesian network,syndrome importance distribution and conditional probability of corresponding nodes respectively of these three syndrome factors.Conclusions:1.The combination of western medicine name and TCM name is common in TCM clinical;so disease,syndrome types and syndromes are.The disease of TCM and Western medicine is relatively concentrated,and the dominant disease of TCM is obvious;TCM syndrome types are various,and the common syndrome types are concentrated;syndrome factor diagnosiscould make clinical easier to be mastered;The same syndromes of different diseases are relatively concentrated.2.TCM syndrome types are mainly composed of 1-4 syndrome elements;Clinical syndrome is mainly composed of multiple syndromes,and simple syndromes are rare;The combination rules excavated conform to TCM clinical thinking and can guide clinical diagnosis and treatment.3.The diagnostic regularity of bayesian network mining is in good agreement with the clinical dialectical thinking of TCM,and we can dig out the hidden information under the large data at the same time.Although it does not fully reflect the dialectical thinking of TCM,it has certain advantages for mining TCM syndrome differentiation information,which can provide a matching paradigm for syndrome factor differentiation model in TCM health management system.