递推最小二乘方法用于形变时变模型的研究

RECURSIVE LEAST SQUARE METHOD APPLIED IN STUDYING DEFORMATION AND TIME-VARYING MODEL

递推最小二乘方法是以最小二乘算法为基础所建立的关于参数估计的一种递推式方法。它具有最小二乘估计所具有的一切特点,即参数估计具有无偏性、有效性和一致性。适合于形变动态模型的追踪。 本文研究了楚雄台水管观测和牛口峪水准观测模型随时间的变化。并讨论了模型参数的变化与地震孕育、发生之间可能存在的关系,发现地震前形变模型参数是有异常变化的。 本文亦对形变模型的物理意义进行了,必要的讨论。

The recursive least square method is a recurrence method established based on the least square algorithm for the parametric estimation. This method has all the characteristics which the least square estimation possesses, namely, the parametric estimation is of unbiasedness, effectiveness and uniformity. It is suitable for the tracing of the deformation-dynamic model. This paper studied the variations of the water-tube observation model at Chuxiong station and the levelling observation model at Niukouyu station with time. The paper also discussed the possible relation between the variation of the model parameters and the preparedness and occurrence of earthquakes. It is found that the parameters of deformation model show abnormal variation before arthquakes. The paper also discussed the physic significance of deformation model.