[SciPy-User] ier-integer in optimize.leastsq

Andrew Giessel andrew_giessel@hms.harvard....
Mon Mar 26 11:10:32 CDT 2012


In other words, it is a function of the precision of the number type
(float) you use in your fitting routines (specific to your
computer architecture).

hth,

ag

On Mon, Mar 26, 2012 at 09:54, <Philip_Bransford@vrtx.com> wrote:

> 1.49012e-08= numpy.sqrt(numpy.finfo(float).eps)
>
>
>
> From:        "Johannes Radinger" <JRadinger@gmx.at>
> To:        scipy-user@scipy.org
> Date:        03/26/2012 06:11 AM
> Subject:        [SciPy-User] ier-integer in optimize.leastsq
> Sent by:        scipy-user-bounces@scipy.org
> ------------------------------
>
>
>
> Hi,
>
> Some months ago I started already this topic,... Now, while writing my
> manuskript this topic comes up again. I am using the optimize.leastsq
> function
> and would like to describe my results especielly the "ier-level".
>
> If ier in my result is 1, then the solution which was found ensures a
> ftol-"quality", resp. the sum of squares for the relative errors are below
> the ftol value?
> As I didn't set any ftol value, the default value is used. Therefore, in
> the case of optimization
> results with ier=1, is it possible to state:
>
> "For all fitted solutions the relative errors in the sum of squares are
> below the desired standard value of 1.49012e-08" ???
>
> Where does this standard value come from?
>
> Best regards,
>
> Johannes
> --
> NEU: FreePhone 3-fach-Flat mit kostenlosem Smartphone!
>
> Jetzt informieren: http://mobile.1und1.de/?ac=OM.PW.PW003K20328T7073a
> _______________________________________________
> SciPy-User mailing list
> SciPy-User@scipy.org
> http://mail.scipy.org/mailman/listinfo/scipy-user
>
>
>
> _______________________________________________
> SciPy-User mailing list
> SciPy-User@scipy.org
> http://mail.scipy.org/mailman/listinfo/scipy-user
>
>


-- 
Andrew Giessel, PhD

Department of Neurobiology, Harvard Medical School
220 Longwood Ave Boston, MA 02115
ph: 617.432.7971 email: andrew_giessel@hms.harvard.edu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.scipy.org/pipermail/scipy-user/attachments/20120326/0d7b5cfd/attachment-0001.html 


More information about the SciPy-User mailing list