基于信噪比检验的部分奇异值双参数修正估计

发布时间：2018-02-16 02:43

本文关键词： 信噪比复共线性岭估计奇异值双参数　出处：《测绘科学技术学报》2017年04期 　论文类型：期刊论文

【摘要】：从分析LS估计和岭估计的奇异值展开式出发,研究了病态性问题的解算方法。岭估计中岭参数对较大奇异值的修正不能有效降低估计的方差,却增大了估计的偏差,这是导致岭估计有时效果不好的重要原因。首先借鉴TSVD思想,保留展式中较大奇异值所对应的项,然后修正展式中较小奇异值所对应的项。在修正前,利用信噪比检验,将各个参数按受到复共线性危害的大小分为两类,据此作出不同强度的双参数修正,提出了基于信噪比检验的部分奇异值双参数修正估计。实验结果表明,新方法的解算精度较高。
[Abstract]:Based on the analysis of the singular value expansions of LS estimation and ridge estimation, this paper studies the solution method of ill-conditioned problem. The correction of ridge parameter to larger singular value in ridge estimation can not effectively reduce the variance of estimation, but increase the deviation of estimate. This is an important reason for the bad effect of ridge estimation. First, the term corresponding to the larger singular value in the expansion is retained, and then the term corresponding to the smaller singular value in the expansion is corrected. Before the correction, the signal-to-noise ratio (SNR) test is used. Each parameter is divided into two categories according to the magnitude of the complex collinear hazard. Based on the two parameter correction of different intensity, a partial singular value two-parameter correction estimation based on SNR test is proposed. The experimental results show that, The accuracy of the new method is high.
【作者单位】：信息工程大学;
【基金】：国家自然科学基金项目(41174005;41474009) 信息工程大学自主科研课题项目(2016603006)
【分类号】：P207

【相似文献】