Concerned with the training process and accuracy, the LM algorithm is superior to conjugate gradient algorithm and a variable learning rate back propagation (BP) algorithm.
就训练次数与精确度而言，它明显优于共轭梯度法及变学习率的BP算法，适用于系统辨识。
In this paper, a kind of variable metric fast second order nonlinear optimization algorithm is proposed, where a second order interpolating method is used in the optimization of learning rate.
神经网络的辨识采用变尺度二阶快速学习算法，利用二阶插值法来优化搜索学习速率。
应用推荐