Could someone please explain the
answer to this?

Why is the definition of the SVD in
Numerical Recipes different from the
definition in linear algebra books
such as Strang or Leon? Are there two
different ways to formulate the SVD?

SVD:
A = U S Vt
where Vt = V transpose

Numerical Recipes:
U is mxn, S is nxn

Strang, Leon:
U is mxm, S is mxn

Using the NR SVD, how would one
compute the pseudo-inverse?

Thank you for any information.

Linda Seltzer
lseltzer@princeton.edu

i also wondered about it for a long time. now i understand. the definition of the NR is thin svd. if you want to find more information, please read the book "matrix computations". I hope this will help you.