SVD
7 views (last 30 days)
Show older comments
In MATLAB,if we take the svd(X)of a column matrix,we r supposed to get the product of three matrices after decomposition.I want to take the pseudo inverse of this matrix pinv(svd(X)).but after svd I am getting a single number instead a matrix.Pl help me why like this.I am reqd to maultiply another column matrix to this inverse matrix to get the answer. Ex: Force F = pinv(svd(FRF))*A, actually I am expecting pinv(svd(FRF)) should be a 1x3 size and Acceleration matrix A should be of 3x1.Kindly help me to resolve this Shrinivas Gombi
0 Comments
Answers (2)
Walter Roberson
on 13 Jul 2011
svd(X) by itself returns a vector of singular values. The vector will be a column vector of length min(size(X)) . When X is a column vector, min(size(X)) will be 1, so svd(X) will return a scalar.
If you want to be working with matrices, you should probably be using the multi-output version of svd,
[U,S,V] = svd(X);
2 Comments
Walter Roberson
on 14 Jul 2011
[u s v] = svd(FRF(i));
F = pinv(s);
Note: this will be equivalent to
F = zeros(1,length(i));
s = svd(FRF(i));
if s ~= 0; F(1) = 1/s; end;
Krishna Kumar
on 13 Jul 2011
When you do svd of a column matrix, you get 3 matrices [u s v]=svd(X) if X is nx1 vector, u is n*n,s is n*1 and v is 1*1 now if use svd(X) without output args you get 1x1 value, i.e the first entry in s. The remaining entries in 's' are actually zeros. Hope this could help you find a way out.
See Also
Categories
Find more on Surface and Mesh Plots in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!