Abstrakt

The research of digital recognition technology based on bp neural network

Xiaofeng Wang

BP neural network is widely used in nonlinear modeling, pattern recognition and forecasting etc. Based on principles of BP neural network to digit recognition, using several training functions prepared by the MATLAB neural network toolbox, the paper focus on training of the BP neural network, and uses the Smolensk neural networkmodeling to simulate it. The result shows that the BP neural network used for the digit recognition is capable for recognition to some extent, its algorithmoperated easily, and properly choosing relative variables can achieve better recognition effect of the system.