tags:

views:

57

answers:

1

Just read something where it says to decompose a vector, say A.

[U S V]=svds(A,2).

Now, the document says to use the D and P of the decomposed vector in some other function. I guess, D is the diagonal elements S. What is P indicate ?

I guess, its a stupid question but I just cant figure it out.

+1  A: 

One notation that comes to mind is the diagonalization of a real symmetric matrix, usually denoted A = PDP^T. In this case D is a diagnoal matrix of eigenvalues and P can be build from the corresponding eigenvectors

See e.g. http://en.wikipedia.org/wiki/Diagonalizable_matrix

second
Haven't confused with any other decomposition for sure. Document says to use singular value decomposition on A. I used Matlab svds to do that. I believe, then I end up with U S V. Latter the document has written a function, 'SIM' where it says to input D and P of decomposed A. I'm going through the wiki atm.
Tinglin
the name of the decomposed components are just convention, and usv is the standard. unless you can link to the original document, my only advice is to reread from the beginning and look for a definition of p (or svd)
second
The author just replied. He wanted to mean D=S and P=(U*V').
Tinglin

related questions