时域最小二乘配置求定点形变特征曲线与异常识别的应用研究*

(中国地震局第二监测中心,陕西 西安 710054)

最小二乘配置; 定点形变; 异常识别

Application Study of Least Square Collocation in Fixed Point Deformation Characteristic Curve and Anomaly Identification
JIA Peng,ZHANG Xi,TANG Hong-tao,LI Rui-sha

(The Second Monitoring and Application Center,CEA,Xi'an 710054,Shannxi,China)

least square collocation; fixed point deformation; fitting; extrapolation

备注

将最小二乘配置方法的高斯型经验协方差函数模型应用于定点形变特征曲线的异常辅助识别,通过两个不同算例说明最小二乘配置滤波、内插和外推可以进行定点形变曲线的趋势模拟和辅助查找其短期、中长期的异常特征,对于定点形变曲线异常识别有一定的意义。

The Gauss-Function covariance model based on least square collocation was applied in the anomaly assisted identification of fixed point deformation characteristic curves. We used two different examples to show that the filtering,interpolation and extrapolation of least square collocation can be simulated the trend of fixed point deformation curves and assisted searching their anomaly characteristics in short-,medium-and long-term. The least square collocation method has certain significance for identifying the fixed point deformation.