Feedforward neural network is an important network structure of neural computing.It is used widely in engineering optimization.It's error correction algorithm is an important factor of determining the network performance.That is of great significance to data mining,neural network integration,hardware implementation of neural network and distributing parallel processing.This paper is about analysis of a few error correction algorithm.They are compared by examples.The significant conclusions are drawn from it...