Research Of Complex Network Analysis Method Of Pulse Wave

Pulse wave of human life is an important signal containing information on the human condition,the pulse wave of research has been one of the main objects of study of human health related fields.How to find useful physiological information of human body in the pulse wave is the key of the pulse wave research.Existing research methods are mainly looking for a fixed amount over an extended period or the same amount of a given short period of time,such as the amplitude of the frequency domain analysis and the time of cardiac ejection in time domain analysis,for an extended period of pulse wave changes introduced without adequate research.This thesis mainly aimed at the change of the pulse wave in a long time,looking for the body information contained in pulse wave.Changes in pulse wave contains a wealth of information about the human body,is already a complex system.Based on the study of the pulse wave and complex network,this thesis introduces the analysis of complex network into the analysis of pulse wave.The visibility graph and horizontal visibility graphs are applied to the complex network of pulse wave after its appropriate modifications.Both of these methods already have been numerous achievements in the analysis of time series.In the data preprocessing of the complex network analysis of pulse wave,the periodic recognition of the pulse wave is the first step.Existing pulse wave recognition algorithm includes a differential threshold,regional extremes and envelope solutions.These methods have some disadvantages such as the continuous and smooth of the pulse wave,or the large influence of the noise,or the excessive complexity of the computation.In this thesis,the regional extreme value method is improved and the periodic recognition algorithm of pulse wave is obtained.This algorithm has good recognition results in a variety of different situations pulse wave.In this thesis,the complex network of pulse wave is constructed by the periodic sequence of pulse wave.The influence factors of the weight values of the edges in the original time series network method are expanded,so that the pulse wave network contains more pulse wave information.In addition,the network analysis of network levels is carried out,and its non-scale characteristic is found,which provides support for the application of the research results in complex networks.At the same time,the pulse wave network analysis method is applied in the actual experiment,which is used for the identification of different human body state,such as forward and backward movement state recognition,state recognition before and after drinking,etc.The identification results were compared with the frequency domain analysis method,and also analyzed the time complexity of the algorithm.The results show that the complex network on the part of the state of human body recognition effect has advantages such as drinking and state recognition,calculation time complexity also dropped to O(n)level.