最小二乘估计的HOUSEHOLDER变换快速递推算法  

RECURSIVE FAST ALGORITHM FOR LEAST SQUARES ESTIMATION WITH HOUSEHOLDER TRANSFORMATION

作　　者：孟晓风[1] 王行仁[1] 黄俊钦[1] 

机构地区：[1]北京航空航天大学

出　　处：《自动化学报》1994年第1期20-28,共9页Acta Automatica Sinica

摘　　要：本文利用ＨＯＵＳＥＨＯＬＤＥＲ交换（简称Ｈ－变换）推导出最小二乘估计的递推算法和遗忘因子法的快速算法．与现有的最小二乘递推算法相比，本文提出的算法不仅运算量大大减少，而且数值稳定性好，占用内存量少．Based on Householder transformation, two recursive fast algorithms for parameter estimation, i.e., Recursive Least Squares Fast Algorithm and Recursive Forgetting Factor Fast Algorithm, are presented in this paper. As compared with conventional LS recursive algorithms, the proposed algorithms are suitable for the enhancement of numerical stability and the reduction of the number of arithmetic operations.

关 键 词：参数估计 最小二乘法 快速递推算法 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]