na_lec_1

na_lec_1 - 1. Singular Value Decomposition (SVD) Unless...

Info iconThis preview shows pages 1–3. Sign up to view the full content.

View Full Document Right Arrow Icon
1. Singular Value Decomposition (SVD) Unless stated to the contrary, and are integers, and moreover . m n n m 1.1. The matrix , can be decomposed into n m × A T V U A Σ = where , m T T I UU U U = = { } n m i n i diag × = = Σ ,..., 2 , 1 , 0 σ , and . n T T I VV V V = = Moreover (a) The singular values n i i ,..., 2 , 1 , = are the square roots of the eigenvalues of and . T AA A A T (b) The columns of are eigenvectors of . U T AA (c) The columns of are eigenvectors of . V A A T Remark 1.1. The SVD do exist also for the case where n m < . But we will concentrate on the other case. 1.2. The compact form of the SVD is T V U A Σ = where n p is the number of nonzero eigenvalues of , and A A T { } , , ,..., 2 , 1 , 0 , , p T p n p p i m T p m p i diag I V V V I U U U = = = Σ = × × × . This result follows from the block-matrix multiplication rule. 1.3. With p as in 1.2 the SVD can be written in summation form = = p i T i i i 1 v u A where [] m u u u U L 2 1 = and [ ] n v v v V L 2 1 = . 1
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
1.4.1. The pseudoinverse of a diagonal matrix D p i d d d d i n m p ,..., 2 , 1 , 0 ; 0 0 2 1 = = × O D O O is denoted by and given by + D () () () m n p d d d × + = 0 0 1 1 2 1 1 O O O D 1.4.2. The pseudoinverse of is denoted by and given by A + A (a) , or equivalently by m n T × + + Σ = U V A (b) , provided that
Background image of page 2
Image of page 3
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 10

na_lec_1 - 1. Singular Value Decomposition (SVD) Unless...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online