Study On The Methods Of Physiological Signal Prediction With Support Vector Machine

The living creature signal- eye electricity, brain of the conduct and actions typical model electricity, the muscle telecommunication number, contain abundant information about physiology, psychology and pathology. The physiological signal process and real-time forecasting analysis have enabled us to promote the people of health, raise the safety degree , improve the people's living quantity and reduce the trouble about lack of sleep or the tiredness as a result, the signal processing and signal analysis have actively meaning in reality. The eye electricity,the brain electricity,the heart electricity and muscle electricity all belong to the living creature electricity, they imply a large number of the depth information of vigilance and sleep. Therefore, it is an exact research subject that how to valuate and predict the future fettle according to the physiology data which have already obtained the variety of the physiology appearance for a period of foregoing time.This dissertation focuses on the studies of the physiological signal process and forecasting analysis which used the support vector machine.The major work of the dissertation comprising: For constructing the general frame based on the SVM time sequence, the nonlinear dynamics theory and the correlation technique was used to predict the time series for the reconfiguration of phase space. In other hand, According to the signal of EOG and its winking characteristic, some regression estimate about the giving settle tester and the parameter of relevant winking characteristic were investigated. And the same time, the predictive model building and predictive simulation of the affair by hit were also reached a conclusion. Furthermore, the predictive consequence used the scale evaluation was contrasted to the results of the collaborator, Cox, we had finally come to conclusion that the method has better characteristic. In addition, the predictive rumble strips was also contrasted to the hit time, the analysis results show that there is no inevitable causal relation between the personal tiredness and the number of rumble strips about settle tester impact stop.The method of time sequence about the support vector machine was applied...