用改进的非监督学习方法加速前馈神经网络的训练

本语文对非学习方法和前馈神经网络的结构进行了一些简单的改进，用改进后的结构和学习方法寝化前馈神经网络的权值，使前馈神经网络的训练速度和逼近精度显著提高。最后本语文通过仿真实例验证了该方法。

Accelerating the Training of Feedforward Neural Networks Using Improving Unsupervised Learning Principle

In this paper,a little simple improvement was made on unsupervised learning principle and feedforward neural networks configuration.The synaptic weights of the networks is initiated with them.It is advanced notability for the networks learning.At last, the simulation is given to show the properties of the method.