论文部分内容阅读
Focused on various BP algorithms with variable learning rate based on network system error gradient, a modified learning strategy for training non-linear network models is developed with both the incremental and the decremental factors of network learning