一种输入驱动的BP网络高效学习算法

A High-Efficiency Learning Algorithm of Input Driven BP Neural Network Model

在深入研究BP神经网络模型的基础上,从网络互连方式、网络权值初始化以及隐结点的选取等方面,对标准的BP算法作了改进.通过实验。

In order to improve the efficiency of BP neural network model, this paper presents a high efficiency learning algorithm. It adopts input layer driven and optimally initializing weights and automatically selecting the number of hidden nodes. The experiments show that the learning algorithm is reliable and efficient.