论文部分内容阅读
以Sigmoid为传递函数的BP网络在过程系统工程领域已经得到了广泛的应用 ,但是一般的GDR训练算法在极小点附近易发生振荡 ,收敛速度慢。本文提出了人工神经网络M法训练的新途径 ,并且通过不同算例和工业实际数据建模应用证实了M算法的收敛速度大约是GDR算法的 5 - 10倍左右 ,有效地提高了网络训练的速度和训练效率
The BP network with Sigmoid as the transfer function has been widely used in the field of process system engineering. However, the general GDR training algorithm is prone to oscillation near the minimum point and has a slow convergence rate. In this paper, we propose a new way to train M method by artificial neural network, and prove that the convergence rate of M algorithm is about 5 - 10 times that of GDR algorithm through different examples and practical industrial data modeling and effectively improve the training of network Speed and training efficiency