[
https://issues.apache.org/jira/browse/MATH-1008?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13716385#comment-13716385
]
Gilles commented on MATH-1008:
------------------------------
I wish to proceed as I proposed; the move seems adequate according to the first
line of [that page|http://en.wikipedia.org/wiki/Non-linear_least_squares].
If and when someone comes up with implementations of a broader class of
optimizers of which the current "LevenbergMarquardtOptimizer" will be a
subclass, it will be possible to refactor again if another package layout is
deemed more appropriate.
Moreover, doing the move now has practical objectives:
* Provide a testing field for the "fluent API" (as agreed on the "dev" ML)
without cluttering the existing classes.
* Make it possible to further clean up the classes without being encumbered
with compatibility-breaking considerations.
Any technical objection?
> Move the contents of "o.a.c.m.nonlinear.vector.jacobian" to the
> "o.a.c.m.fitting" package
> -----------------------------------------------------------------------------------------
>
> Key: MATH-1008
> URL: https://issues.apache.org/jira/browse/MATH-1008
> Project: Commons Math
> Issue Type: Sub-task
> Reporter: Gilles
> Assignee: Gilles
> Priority: Minor
> Labels: api-change
> Fix For: 3.3
>
>
> The non-linear least-squares optimizers should be considered as "fitting"
> algorithms and thus moved to the "o.a.c.m.fitting" package.
> As agreed on the [dev ML|http://markmail.org/message/3fjvucyz7rax4cyi], the
> new classes will use the "fluent API".
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
