Study On Arterial Pressure Prediction And Realization Of The System Based On Neural Network

The blood pressure and heart disease has become one of the chronic diseases. So the research of the arterial pressure management and prediction is significant. This paper will discuss the research on arterial pressure forecasting based on artificial neural network.It has been proved that as long as the hidden layer has enough neurons, the BP neural network can approximate any nonlinear model. With the dynamic feedback connection, the Elman neural network has the function of associative memory that BP lacks. But some challenges still exist in practical with BP and Elman neural network. In addition, the analysis of time series prediction based on echo state network has been paid extensive attention in recent years. This paper also discusses the mathematical model, learning algorithm and the key parameters of echo state network.In order to monitor the instability of human blood pressure, the research on arterial pressure forecasting becomes more and more important. This paper will discuss deeply about the model structure and characteristic of BP, Elman and echo state network. By using matlab to design three kinds of neural network model, we trained the model and simulated the changes of arterial pressure. Meanwhile, we chose the most suitable one for arterial pressure series prediction according to the prediction accuracy, stability and operation efficiency.Experiments have shown that arterial pressure forecasting based on the echo state neural network has high accuracy, fine adapt ability and stable performance. Meanwhile, a management and prediction of blood pressure system based on ESN was established and had good performance.