Abstract
Neural network has ability of self-studying, self-adapting, fault tolerance and generalization. However, there are some defaults in its basic algorithm, such as low convergence speed, local extremes, and uncertain number of implied layer and implied notes. So there are some limitations in practice. In order to avoid these shortages, the paper solves these problems from two aspects. One is to adopt principle component analysis to select study samples and to make some of them containing more sample characteristics; the other is to train the network by using Levenberg-Marquardt backward propagation algorithm. Finally, an example is used to prove the new method is of high effectiveness and practicality in solving the addressing problem of garbage power generation plants
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have