摘要
随着观测信息的不断丰富,常规的线性模型的不足逐渐凸现,影响了沉降数据分析的精度。据此,为了克服常规沉降数据处理模型的不足提出了动态参数线性回归模型。该模型要求参与建模的数据个数保持不变,利用新观测的数据取代旧数据,获取新的模型参数,进而建立动态参数线性回归模型进行沉降预测。通过实例验证,动态对数模型预测值平均误差率为5.4%,动态多项式模型预测值平均误差率为2.7%;与其对应的常规模型相比平均预测精度分别提高了6.4和2.1个百分点,动态模型的效益十分明显;从建模精度σ2来看,动态对数模型σ2=11.5,动态多项式σ2=2.3,均小于常规模型的σ2。
Linear regression model is one of the most widely used methods in the prediction of embankment settlement. With the data augmented, the disadvantages of the traditional linear model stands out extremely, so the accuracy of prediction is reduced. To overcome deficiencies in the traditional model, dynamic linear regression model is put forward. This model requests that the number of data used to make model is kept unchanged, new data replaced the old data, so the new parameters are gained. Ideal result is gotten through the instance. The average error of Dynamic Model was 5.4% , and the average error of Dynamic Polynomial Model was 2.7%. For the model accuracy σ^2, Logarithm Dynamic Logarithm Model was σ^2— = 11.5 and Dynamic Polynomial Model was σ^2- = 2.3, both of them all below general model σ^2-.
出处
《地矿测绘》
2008年第1期12-15,共4页
Surveying and Mapping of Geology and Mineral Resources
关键词
沉降预测
线性回归
动态参数模型
embankment settlement prediction
linear regression
dynamic parameter model
