Re: Linear regression without intercept in NCL

From: Dennis Shea <shea_at_nyahnyahspammersnyahnyah>
Date: Sat Jan 12 2013 - 19:05:47 MST

[1] There is no function "for a linear regression without
     intercept, that is, when x=0, y=0." In fact, this is not
     advisable. See:
http://www.talkstats.com/showthread.php/19521-Linear-Regression-Without-intercept

     Note: if you have further questions on this topic, please ask a
     statistical web site. ncl-talk is for questions about the language.

[2] Please search the list of functions
     http://www.ncl.ucar.edu/Document/Functions/list_alpha.shtml

     There are functions for RSMD (same as RMSE) and standard deviation.
     The "coefficient of determination"

[3] There is no explicit function of "coefficient of determination."
     You can calculate it yourself.

[4] There are several functions that calculate the linear correlation
     coefficient and regression (r). The square of r [r^2] is a measure
     of the the total variation in y can be explained by the linear
     relationship between x and y.
------------

Send statistical questions to the R language web site or to a statistics
WWW.

On 1/4/13 9:59 AM, wuhui wrote:
> Hi all in mail-list,
>
> I calculated the linear regression coefficient using build-in
> function regline(), and recognized that the function regline()
> is for a linear regression with a constant (i.e. intercept).
> So, I wonder that is there any function/script for a linear
> regression without intercept, that is, when x=0, y=0?
> Further more, are there some functions/scripts which calculate
> common statistical errors of linear regression, for example,
> RMSE/D (Root-Mean-Square Error/Deviation), R^2 (Coefficient of
> determination)?
>
> Thanks in advance!
>
> Best regards
> Hui Wu
> 05-JAN-2013
>
>
>
> _______________________________________________
> ncl-talk mailing list
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>
_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Sat Jan 12 19:05:57 2013

This archive was generated by hypermail 2.1.8 : Mon Jan 14 2013 - 14:31:38 MST