Data of the effective cases treated by Chinese medicine were collected from the databases of CNKI,Wanfang and VIP for the classification of traditional Chinese medicine syndrome patterns of urticaria based on the models of BP(back propagation)neural network and support vector machine(SVM),and then the accuracy of BP neural network syndrome classifier and SVM syndrome classifier was compared. The results showed that BP neural network syndrome classifier had the accuracy of 83.13% and SVM syndrome classifier had the accuracy of 92.3%,indicating that the latter one had the higher accuracy. Both of BP neural network syndrome classifier and SVM syndrome classifier can achieve satisfactory effect for the syndrome classification of urticaria. For the research is just in the exploration stage of theoretical model,large sample size and intelligent optimization algorithm will be needed to increase the accuracy of BP neural network syndrome classifier and SVM syndrome classifier.
Journal of Guangzhou University of Traditional Chinese Medicine
BP neural network
support vector machine