Re: [Help-gsl] non-square QR decomposition

[Brian Gough]

Martin Jansche writes:
 > I'm trying to solve a least squares problem involving a linear system
 > of the form Ax=b, where A is an n*k matrix, x is a k-dimensional
 > vector, and b is an n-dimensional vector. QR decomposition works fine,
 > but gsl_linalg_QR_solve() complains that its first argument (the QR
 > matrix in its internal storage format) must be square. Given that the
 > documentation talks about general rectangular matrices (unlike
 > Numerical Recipes code, which is restricted to square matrices), I
 > expected this to work. What am I missing?  I know about gsl_multifit,
 > but I need something more general.

There is a separate function for the non-square case ("lssolve", I
think).  See the GSL manual for details.

-- 
Brian Gough

Network Theory Ltd,
Publishing Free Software Manuals --- http://www.network-theory.co.uk/