长江科学院院报 ›› 2003, Vol. 20 ›› Issue (2): 50-53.

• 工程监测 • 上一篇    下一篇

大坝神经网络预报模型与大坝回归预报模型的比较

 吴云芳, 李珍照   

  1. 武汉大学 水利水电学院, 湖北 武汉430072
  • 出版日期:2003-04-01 发布日期:2012-03-05

Comparison between BP neural network model and statistical regression model of dam

 WU  Yun-Fang, LI  Zhen-Zhao   

  • Online:2003-04-01 Published:2012-03-05

摘要: 了更好地应用BP神经网络预报模型和回归预报模型,通过实例从模型原理、模型建立和模型应用三方面对两种预报模型进行了比较。在模型应用方面,主要从因子相关性、资料系列的长短、自变量因子的不同组合等角度对两种模型的影响进行了分析。结果证明:BP模型在大坝监测值预报中是一种性能优异的模型。当需对大坝的监测数据作分解和解释时,则适宜采用统计回归模型。

关键词: BP神经网络模型, 统计回归模型, 模型比较

Abstract: To make full use of BP neural network model and statistical regression model of dam, through illustrations a comparison between them is made in three aspects, i.e. from model principle, model formation and their application respectively. Especially in the application of model, the paper emphatically analyses the influences of factor correlation, data time-interval, and different combinations of independent factors on these two kinds of models. The results show that the BP neural network model is an excellent one in prediction monitoring data of dam, while the statistical regression model is suitable for analyzing and explaining the obtained monitoring data.