请输入您要查询的字词:

 

单词 ProofOfExistenceAndUniquenessOfSingularValueDecomposition
释义

proof of existence and uniqueness of singular value decomposition


Proof of existence and uniqueness of SVDFernando Sanz Gamiz

Proof.

To prove existence of the SVD, we isolate the direction of thelargest action of Am×n, and then proceed by inductionMathworldPlanetmath on the dimensionMathworldPlanetmathPlanetmathPlanetmathof A. We will denote hermitian conjugation by T. Norms forvectors in n will be the usual euclideanPlanetmathPlanetmath 2-norm =2 and for matrix the induced by norm of vectors.

Let σ1=A. By a compactness argument, there must bevectors v1n,u1*m with v1=1,u1*=σ1 and u1*=Av1. Normalize u1* bysetting u1=u1*/u1* and consider any extensionsPlanetmathPlanetmath ofv1 to an orthonormal basis {vi} of n and of u1 to anorthonormal basis {uj} of m; let U1 and V1 denotethe unitary matricesMathworldPlanetmath with columns {vi} and {uj}respectively. Then we have

U1TAV1=S=(σ1wT0B)

where 0 is a column vectorMathworldPlanetmath of dimension m-1, wT isa row vector of dimension n-1, and B is a matrix of dimension(m-1)×(n-1). Now,

(σ1wT0B)(σ1w)σ12+w2=(σ12+w2)1/2(σ1w)

so that S(σ12+w2)1/2. But U1and V1 are unitary matrix, hence S=σ1; ittherefore implies w=0.

If n=1 or m=1 we are done. Otherwise the submatrixMathworldPlanetmath Bdescribes the action of A on the subspaceMathworldPlanetmathPlanetmath orthogonalMathworldPlanetmathPlanetmath to v1. Bythe induction hypothesis B has an SVD B=U2Σ2V2T. Now itis easily verified that

A=U1(100U2)(σ100Σ2)(100V2)TV1T

is an SVD of A. completing the proof of existence.

For the uniqueness let A=UΣVT a SVD for A and let eidenote the i-th, i=1min(m,n) vector of the canonical base of n. As U andV are unitary, Aei=σi2, so eachσi is uniquely determined.

随便看

 

数学辞典收录了18232条数学词条,基本涵盖了常用数学知识及数学英语单词词组的翻译及用法,是数学学习的有利工具。

 

Copyright © 2000-2023 Newdu.com.com All Rights Reserved
更新时间:2025/5/4 5:07:20