中文说明：该代码是基于大数据的逐层学习的受限玻尔兹曼机程序，通过受限玻尔兹曼机的逐层学习提取特征，并进行解码，来得到重构结构，然后使得重构结果尽可能接近输入值，因此运用反向传播算法对参数进行调优，是使结果达到最好。
English Description：
The code was based on the data of layer-by-layer restricted Boltzmann machine learning procedures through layer-by-layer for restricted Boltzmann machine learning characteristics extracted and decoded to get reconstructed structure, which then allows refactoring result as close as possible to enter a value, therefore the use of the back-propagation algorithm to tune parameters, is to make the results best.