[ncl-talk] confidence interval of mean response in regline_stats function

[Dennis Shea]

The function code is located in the "contributed.ncl" library.
Search for   'regline_stats'
The function '*cdft_t'*
<http://www.ncl.ucar.edu/Document/Functions/Built-in/cdft_t.shtml>  from
the DCDFLIB <http://www.netlib.org/random/> (Double precision Cumulative
Distribution Function LIBrary) is used to calculate the -t-values. You will
have to look at that code to see what they do.

Good Luck




On Mon, Nov 16, 2020 at 7:08 AM 黄涵 via ncl-talk <ncl-talk at mailman.ucar.edu>
wrote:

>
> Hi All,
>
> I would like to know whether any of you know how regline_stats function
> calculates the confidence interval (CI) of the mean response (attributes:
> YMR025 and YMR975).
>
> As this function only returns back the CI of those points used in the
> linear regression model,  if I want to get the CI of a x-value that is not
> used in the regression model, I have to calculate it by myself.  I followed
> the formula in the attachment to calculate the CI, but there are some
> differences between my result and the one provided by this NCL function
> (see attached plot, especially around the x_mean region).
>
> Thanks very much.
>
> Han
> _______________________________________________
> ncl-talk mailing list
> ncl-talk at mailman.ucar.edu
> List instructions, subscriber options, unsubscribe:
> https://mailman.ucar.edu/mailman/listinfo/ncl-talk
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.ucar.edu/pipermail/ncl-talk/attachments/20201116/52666902/attachment.html>