Data-based Study On The Therapy For Palpitation Phlegm-fire Disturbing Heart

Objective: Based on the medical case record of ancient physician and the collected typical clinical cases, ancient and modern medical case database of Palpitation is established. Data exploration method is established based on the decision tree and neural network to interpret such information as differentiation of syndrome, syndrome complex,etc. of palpitation phlegm-fire disturbing heart in medical case record of ancient physician and the collected typical clinical cases, to find the special features and significance of TCM four diagnostic methods in treating palpitation phlegm-fire disturbing heart and to construct the syndrome discrimination model of palpitation and pyrophlegmatic mania. The interpretation of prescription and rule of medication with the inheritance of traditional Chinese medicine auxiliary system( V2.5) is instructive for clinical medication of treating palpitation phlegm-fire disturbing heart.Method: 1. Establishment of medical cases database. Respectively establish syndrome database and the database of Chinese Herb representation through collecting medical case record of ancient physician literature and outpatient data. 2. Normalization of medical cases. Standardize the TCM four diagnostic information and prescription. 3.Quantitative assignment of medical cases. Assign the normalized information with computer-sensible two-valued variable 0 and 1. The occurred syndrome is assigned as 1and the absent is assigned as 0. 4. Establishment of intelligent discrimination model. Using CHAID、CRT、QUEST of SPSS 20.0 and SPSS Modeler 14.1 C5.0 algorithm to analyse syndromes and build the decision tree model, which will provide evidence for the objective study of the syndrome of palpitation, pyrophlegmatic phlegm-fire disturbing heart.Moreover, Build syndrome model with multi-layer perception neural network radial basis function neural network(RBF) neural network of SPSS 20.0. 5. Exploration of prescription rules. Calculate the frequency of medicine through analysing the data of prescription of with palpitation phlegm-fire disturbing heart with the inheritance of traditional Chinese medicine auxiliary system(V2.5)so as to analyse its four characters(FC), five tastes(FT)and channel distributions of herbs, herbal medicines in a prescription, amount, the number and percentage of each medicine. Select couplet medicines in refular use with rules of prescription and do the association analysis. Using the software of “Improved mutual information method” to analyse of degree of relation among different medicines in the prescription. Select the core complex with three to four kinds of medicines using the method of complex system entropy clustering, and analyse the possible new prescription using new prescription analysis function to further instruct clinic medicine.Result:1.Build database of the syndrome and prescription characteristics of palpitation and pyrophlegmatic phlegm-fire disturbing heart. 2. Build diagnosis model of palpitation phlegm-fire disturbing heart with CHAID、CRT、QUEST and C5.0 algorithm.Test the above four models with 10 times cross validation techniques. The judgment accuracies of C5.0、CRT、CHAI and QUEST are respectively 94.1%、92.5%、92.0% and87.4%; specificities are respectively 91.3% 、 91.3% 、 90.1% and 88.8%; sensitivities are respectively 96.9% 、 96.2% 、 92.5% and 85.0%; build syndrome diagnosis model with multi-layer perception neural network radial basis function neural network(RBF) neural network with respective accuracies 93.2% and 97.3%. All results conform to the requirement of statistics. 3. Through analyzing the regular prescription and medicine frequency of palpitation phlegm-fire disturbing heart using the inheritance of traditional Chinese medicine auxiliary system(V2.5), 46 herbs are found being used for over more than twenty times most of which are sweet, bitter and xin, and channel tropism of the first six herbs are spleen, lung, heart, stomach, liver and kidney. Moreover, core complex and new prescription are statistically analysed with improved mutual information method and complex system entropy clustering.Conclusion: With the establishment of the databases of the syndromes and prescription characteristics of palpitation and pyrophlegmatic mania, the study analysed the syndrome distribution regularity of palpitation phlegm-fire disturbing heart using the method of decision tree and neural network; and the study also analysed the herb characteristics of the prescription of palpitation phlegm-fire disturbing heart using the inheritance of traditional Chinese medicine auxiliary system(V2.5)and built standardsystem of syndromes and prescriptions to further instruct clinical treatment.