Least Squares of Symmetric Positive Semidefinite Matrices

1.3k Views Asked by At

What's the best (in terms of computation time and numerical robustness) way to find the least squares solution of $$Ax = b$$ if $A$ is symmetric and positive semi-definite?

If $A$ were symmetric and positive definite, the Cholesky decomposition would seem to be the way to go to build the inverse. But that doesn't work for positive semidefinite cases.

Performing a spectral decomposition to get $A = Q \Lambda Q^T$, and then forming the pseudoinverse $A^+ = Q \Lambda^+ Q^T$ to get $x = A^+ b$ works, but I'm not convinced it's the most efficient.

1

There are 1 best solutions below

1
On BEST ANSWER

According to MATLAB's mldivide():

enter image description here

If the Cholesky Decomposition doesn't work you should use LDL Decomposition.