The application research of improving neural network algorithm in the grain monitoring

In allusion to the insufficiencies such as knowledge acquirement, reasoning ability and self-learning ability, the paper applies neural network into expert system, combines with the system of measurement and control for grain storage, and puts forward an improved BP algorithm. This algorithm does not need the prior hypothesis model and has a good compatibility to the complete and noisy information; it also can solve the non-linear problem well. This new algorithm has coordinated contradictions between learning efficiency and convergence rate and improved skilled speed and convergence rate. From the results of experiment, we can see that the new algorithm has some advantages, such as quickly, validity and practicability.