[ncl-talk] reg_multlin_stats() help page

[Mary Haley]

Hi Will,

I don't know if Dennis responded to you offline, but we have fixed and
updated the reg_multlin_stats documentation page to hopefully make things
more clear.

Please see:

http://www.ncl.ucar.edu/Document/Functions/Contributed/reg_multlin_stats.shtml

--Mary



On Mon, Aug 25, 2014 at 9:21 PM, Will Hobbs <Will.Hobbs at utas.edu.au> wrote:

>  Hi Guys
>
>  I've noticed a few issues on the help file for reg_multlin_stats(),
> which looks a lot like that for reg_multlin() (except for the examples),
> even though the behaviour is different.
>
>  Arguments : there's no listing for the 'opt' argument, but looking at
> the examples this seems to affect the behaviour (i.e. whether or not to
> print the ANOVA results)
>
>  Description: The description includes a dummy variable in the x-array,
> as for reg_multlin. However, when I include this I get an inverse_matrix()
> error (missing values are not allowed), and the stderr, pval and tval
> attributes are set to missing. I also get M+2 coefficients, with the first
> coeff. being always zero.
>
>  If I eliminate the dummy variable from the x-input, these issues go away
> and I get M+1 coeffs.
>
>  This is all consistent with the behaviour described in the examples, but
> is NOT the same as the behaviour in the function 'Description'
>
>  Hope that helps
>
>  Will
>
> _______________________________________________
> ncl-talk mailing list
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20140827/7170e46c/attachment.html