Re: Implementation of mldivide?

[CdeMills]

Jordi Gutiérrez Hermoso-2 wrote:
> 
> 
> I'm not sure what you mean by "analytically". The formula you wrote
> down for the inverse of L is certainly not used by the backslash
> operator. Ignoring that word, the answer is a little complicated,
> because the implementation of the backslash operator is spread out
> over several source files depending on the inputs. You may find these
> answers from the manual more illuminating:
> 
>     
> http://www.gnu.org/software/octave/doc/interpreter/Techniques-Used-for-Linear-Algebra.html#Techniques-Used-for-Linear-Algebra
>     
> http://www.gnu.org/software/octave/doc/interpreter/Sparse-Linear-Algebra.html#Sparse-Linear-Algebra
> 
> 
With B = inv(L.'*L)*L.'*y, 
I can write
dBdZ =
-(inv(L.'*L)*(dLdZ.'*L+L.'*dLdZ)*inv(L.'*L)*L.'*y)+inv(L.'*L)*dLdZ.'*y
i.e. an explicit formula where the derivates dLdZ are apparent. 

I should have made clear that the number of lines of L is >> number of
columns, and that the columns are linearly independent. The purpose is to
use noise propagation technique to infer the noise on B caused by the noise
on Z. This way, I need to be able to express the solution in terms of
operations whose derivatives can be computed.

Regards

Pascal 

--
View this message in context: 
http://octave.1599824.n4.nabble.com/Implementation-of-mldivide-tp3695146p3696737.html
Sent from the Octave - General mailing list archive at Nabble.com.