Research On Human Pulse Data Analysis Method Based On Random Matrix Theory

Random matrix theory (RMT) is an important research field in the Probability and Statistics. It is one of the important ways to study the characteristics of complex systems using random matrix theory. By analyzing the energy spectrum distribution and the eigenvalue distribution of the complex system, characteristics of the correlation behavior in the real data are revealed, and thus the structure and properties of the complex system are studied and analyzed.Chinese medicine believes that rhythm, intensity, shape and other aspects of information the pulse wave shows can reflect the human's blood flow characteristics of their cardiovascular system and many physiological and pathological aspects of nervous system. Previous pulse analysis methods should be combined with medical knowledge,the requirements of common sense and data accuracy are usually more higher.Using random matrix theory to deal with the physiological and pathological information within pulse wave is of important value.This paper will deal with the pulse data according to the method of data modeling and analysis in the random matrix theory (RMT), It will use the single ring law to process the pulse data of the human body and get the corresponding state. Roughly it is divided into four parts:1.The feasibility of dealing with pulse data is analyzed combining the random matrix theory and the knowledge of pulse.Some special matrix is simulated and analyzed.2.The pulse sensor developed by Feyi Technology Co., Ltd. was used as the pulse collection device, and the data collection program was designed. Pulse data were collected from dozens of college students and cardiovascular patients in Xiyuan Hospital of Chinese Academy of Sciences.3.The measured pulse data was preprocessed and modeled. Then the data were processed according to a certain method to get the spicific spectrum distribution study the gradual behavior or the deviation behavior of the empirical spectrum distribution.4.Combining four-quadrant mode, we choose the data in different quadrants for longitudinal and lateral comparison, analyzing and comparing the empirical spectrum distribution, and get the corresponding relationship between state and pulse.