The Research Of Therapy Regularity Of Phlegm-dampness Syndrome Of Hypertension Based On Data Mining

Objectives: Under the guidance of TCM theories, essential hypertension of phlegm dampness related to ancient and modern medical and periodical literature and clinical collection of related records as the research object, based on database and data mining technology, using decision tree CRT, CHAID, C5.0, QUEST algorithm and neural network method to establish hypertension phlegm dampness syndrome a mathematical model, by using the frequency analysis and association rules mining algorithm of prescription drug laws, looking for essential hypertension phlegm dampness should permit combination and drug compatibility rules, based on essential hypertension phlegm dampness syndrome and treatment rules, provide the basis for clinical prescription medication.Methods: 1.Data collection. Collect information about the hypertension of ancient and modern medical records of literature and clinic data, the establishment of essential hypertension phlegm dampness and prescription databases.2.Sorting data. Symptoms and medicine on data collected for the standardization of language processing and quantitative evaluation.To the standardization of information using binary variables 0, 1 assignment, the assignment 1 cases syndromes elements, does not appear and the lack of such as data assignment of 0.3. Syndromes on hypertension medical record using SPSS20.0 software(the four diagnostic information) frequency analysis, C5.0, CRT, CHAID and QUEST methods to select on essential hypertension phlegm dampness and non phlegm significance in differentiating congestion syndrome information in theindex, hypertension of phlegm dampness optimal discriminant model, so as to provide basis for the objective study. 4.Hypertension of phlegm dampness optimal discriminant model is established by using SPSS 20.0 multilayer perceptron neural network and radial basis function(RBF) neural network.5.The establishment of hypertension of phlegm dampness prescription information database, input the TCM inheritance system(V2.0.1), the use of complex system improved entropy clustering, mutual information method, unsupervised hierarchical clustering of the entropy unsupervised data mining methods, exploring the law of prescriptions of hypertension of phlegm dampness.Results: 1. Established hypertension syndrome information database and prescription information database. 2. CHAID, CRT, QUEST and C5.0 algorithm is adopted to establish the hypertension of phlegm dampness diagnosis model and the accuracy were 82.9%, 92.1%, 92.4%, 93.74%.the common diagnostic information are head dizzy,Vomiting, Phlegm and saliva,White and greasy fur,Further combined with such as chest stuffiness, fullness, heaviness and other symptoms of the formation of a good combination of phlegm dampness syndrome four diagnostic informationlevel discrimination model.3. Using decision tree and neural network method of syndrome diagnosis discriminant model essential hypertension phlegm dampness and the accuracy were 91..4%, 90.4%, such as head, chest stuffiness, darkened phlegm saliva, vomit, white and greasy fur were the best diagnostic model. 4. The use of "common prescription and drug TCM inheritance system" on essential hypertension phlegm dampness were analyzed, and found the prescription drug laws, derived hypertension phlegm dampness rule is to spleen phlegm, calming liver wind mainly. 5. Through the analysis of two typical cases, obtains the data mining results of clinical diagnosis and treatment of hypertension of phlegm dampness has great guiding value.