[jira] [Commented] (MATH-664) Replace "LUDecompostionImpl" with "QRDecompositionImpl" in "AbstractLeastSquaresOptimizer"

["Gilles (Commented) (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/MATH-664?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13183293#comment-13183293 ] 

Gilles commented on MATH-664:
-----------------------------

In the current design, it is difficult to allow a custom algorithm because the constructor of a decomposition class takes the matrix as an argument, while the Jacobian matrix is constructed internally (which would be _after_ the instantiation of the decomposition class if created by the caller).

A clean way would be to have a new "Decomposition" interface, with a {{decompose(RealMatrix m)}} method. That would entail moving code that is currently in the decomposition classes' constructors over to implementations of that new method.
Since that's unlikely to occur before 3.0, I'll just change the internal algorithm to QR, for now.

                
> Replace "LUDecompostionImpl" with "QRDecompositionImpl" in "AbstractLeastSquaresOptimizer"
> ------------------------------------------------------------------------------------------
>
>                 Key: MATH-664
>                 URL: https://issues.apache.org/jira/browse/MATH-664
>             Project: Commons Math
>          Issue Type: Improvement
>            Reporter: Gilles
>            Assignee: Gilles
>            Priority: Minor
>             Fix For: 3.0
>
>
> In some cases, the "getCovariances()" method throws a "SingularMatrixException". This can be avoided by using "QR" instead of "LU" decomposition.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira