The Research Of BP Neural Network Based On Genetic Algorithm In Pulse Recognition

Pulse diagnosis is an important part of traditional Chinese medicine.It has the indispensable contribution to the development of Chinese medicine.However,the traditional pulse is determined by pulse way,which makes the lack of objective standard pulse identification,affecting its accuracy and feasibility.Therefore,the objective study of pulse diagnosis will be of great significance to traditional Chinese medicine.In this paper,we mainly study and discuss the processing,analysis and identification of four pulse signals of veins,veins,veins and veins.In this paper,the wavelet transform is used to reflect the low frequency and high frequency in the frequency domain.The wavelet decomposition and reconstruction method is used to remove the baseline drift of the pulse wave and eliminate the interference caused by the respiration.At the same time,we obtain the characteristic information from the energy distribution of the 8 scale wavelet decomposition by using the method of wavelet scale decomposition.At the same time,the frequency domain characteristic information of the pulse wave spectrum is obtained by the method of frequency domain analysis.The above feature information is used as the input of the neural network.We obtained the 4 dimensional frequency domain and 6 dimensional wavelet scale energy feature information by extracting the characteristics of the pulse wave.In the process of pulse recognition,we used the improved BP algorithm to identify the frequency domain and wavelet scale energy characteristics respectively firstly.A three-layer BP neural network was established,and 40 samples were trained.The remaining 40 samples were predicted to obtain 75% and 80% classification accuracy respectively.Finally,two characteristic information are combined to form a 10 dimensional eigenvector.We used BP algorithm and the improved BP algorithm based on genetic algorithm to train and predict the above samples,then obtained the classification accuracy of 90% and 95% respectively.Moreover,it is proved that the improved BP neural network based on genetic algorithm has very superior performance in convergence speed and accuracy.Finally,this method is used as the core algorithm of pulse recognition in pulse health monitoring system.Finally,we used the above pulse recognition algorithm research and the previous pulse wave prediction of blood pressure research results to achieve a health care system product by combining with the current development of mobile Internet and intelligent hardware.It makes realtime,portable and intelligent health monitoring step by step.