Artificial neural network based modeling of glucose metabolism

Abstract

Neural network the number of hidden neurons for the network performance has a significant impact, usually for a specific problem, there is no way to determine a certain level in the end should be hidden together the number of neurons, the general test Way through many experiments to achieve the desired effect. The improved BP algorithm, the establishment of the BP neural network diagnostic model, tested its correct diagnosis was 100%, BP model diagnostic accuracy was 95.39%. The results show that the BP neural network suitable for solving the complex problem of disease diagnosis. © 2011 Springer-Verlag Berlin Heidelberg.