基于矩阵奇异值分解的空间直线拟合  

Spatial Linear Fitting Based on Singular Value Decomposition of the Matrix

作　　者：李静 LI Jing(Department of Elementary Education,Army Engineering University of PLA,Nanjing 211106,China)

机构地区：[1]陆军工程大学基础部,南京211106

出　　处：《大学数学》2019年第2期5-8,共4页College Mathematics

基　　金：国家自然科学基金面上项目(11372354);陆军工程大学教育教学课题(GJ1811080)

摘　　要：讨论空间直线拟合问题,先对已知数据进行中心化处理得到矩阵,对此矩阵进行奇异值分解,直线方程中的系数可由相应的列向量确定.该拟合过程不需迭代,简单易行,结果也比原文更接近真实值.The spatial linear fitting is discussed in the paper.A matrix is obtained by data centralization.The singular value decomposition is conducted on the matrix.The coefficients of line equation can be determined by the column vectors from SVD.Without the iteration,the fitting process is simple and the result is closer to the actual value than the original.

关 键 词：线性方程组 直线拟合 广义逆 奇异值分解 

分 类 号：O151.21[理学—数学]