回归神经网络预测模型归一化方法分析 |
| |
引用本文: | 滕明鑫.回归神经网络预测模型归一化方法分析[J].数字社区&智能家居,2014(3):1508-1510. |
| |
作者姓名: | 滕明鑫 |
| |
作者单位: | 重庆市轨道交通集团有限公司,重庆400042 |
| |
摘 要: | 该文运用几种常见数据归一化方法分别对自回归神经网络动态预测模型的预测性能进行分析,结果说明不同数据归一化处理对模型的性能影响非常明显,运用最大运算法进行归一化处理要优于其它几种常见归一化方法。
|
关 键 词: | 回归神经网络 时间序列 数据预测 归一化方法 |
The Analysis of Normalization Method of Recurrent Neural Network Prediction Model |
| |
Affiliation: | TENG Min-xin (Chongqing Metro Group Co. Ltd.,Chongqing 400042, China) |
| |
Abstract: | This article is to analysis the performance of Recurrent Neural Network prediction mode when it uses several com-mon data normalization method, result show that the different data normalization is very obvious influence on the performance of model,and the normalization method by maximum operation is better than the other common normalization method. |
| |
Keywords: | recurrent neural network time series data prediction normalization method |
本文献已被 维普 等数据库收录! |