基于独立分量分析的拱坝位移预测模型

在逐步回归模型中,有些分量间存在比较强的相关性,如果直接利用线性回归方法处理,就会特别容易出现病态问题,这样得到的回归模型不合理。为了解决这一问题,引入独立分量分析方法(ICA),并运用基于负熵极大化的独立性准则的快速独立分量分析方法(FastICA)分离出独立分量,建立了某拱坝的安全监测位移预测模型;同时为了表明结果准确性,利用逐步回归方法建立了逐步回归模型。结果表明:独立分量回归方法(ICR)相比逐步回归方法可以建立一个比较准确的回归模型,由于独立分量回归分析方法没有剔除不显著因子,考虑了各个因子对因变量的影响,从而更加符合实际情况。

Arch dam displacement prediction model based on independent component analysis

In the stepwise regression model, there are strong correlations between some of the components. If the linear regression method is adopted, the regression ill-posed problem will be very likely to occur, so the regression model obtained will be unreasonable. In order to solve this problem, independent component analysis (ICA) is introduced and the displacement prediction model of a arch dam is established using the fast independent component analysis (FastICA ) method based on the principle of independence of maximum entropy. In order to show the accuracy of the results, the stepwise regression model is established by stepwise regression method at the same time. The results show that the independent component regression method (ICR) can establish a more accurate regression model than the stepwise regression method. The reason is that the independent component regression analysis method does not eliminate the non-significant factors, which takes into account the effect of each factor on the dependent variable, so it is more in line with the actual situation.