一种基于改进型递归神经网络的数据校核方法

A Data Rectification Method Based on Modified Recurrent Neural Networks

对于无法得到数学模型的过程 ,我们提出了基于改进型自递归神经网络的动态过程数据校核方法 ,并基于递归网络结构推导出一种BP(BackPropagation—反向传播 )训练方法 .由此方法而得到的数据校验结果更准确无误 .

For the dynamic processes whose models can not be obtained,a data rectification method by modified recurrent neural networks has been presented .The derivation of the back propagation algorithm for these recurrent networks has been given .The data reconciliation method by modified recurrent network is shown to provide superior results.