【摘要】 针对时间序列常用递推算法建模存在的缺点,提出一种适用于预报模型参数估算的新算法,应用动态系统模型参数随机化原理,把形变模型噪声归化到形变预报模型参数上,采用多层递阶分析方法建立形变预报模型。用天马台水准资料和厦门台水平摆资料进行的实例分析表明,新算法提高了形变预报的准确度。更多还原

【Abstract】 In view of the shortcoming that the recursion algorithm is usually used to model time series, a new algorithm suitable for the forecast model parameter’s estimatation is proposed,which naturalizes the deformation model noise to the deformation forecast model parameter based on the principle of dynamic system model parameter randomisation,and uses the multi-layer hierarchical analysis method to establish deformation forecast model. The analysis of the Tianma and Xiamen station’s data indicates that the deformation forecast precision is improved obviously.更多还原