Re: r-Squared Question

On 14 Jul 2005 21:40:57 -0700, "Reef Fish"
<Large_Nassau_Grouper@xxxxxxxxx> wrote:
[ snip, much]

[Jerry]> >

> > *I* calculated the residuals: Y-Yhat
> >
> > ResSS is the sum of their (residuals) squares = 85. However, TSS =
> > Sum[(Y-105.5)^2] is only 82.5.
> >
> > I plug those numbers into 1-ResSS/TotSS and get -0.03. Do you get
> > something different?

RF >
> No, I didn't even look at your data! :-) I was looking ONLY
> at the definitions you gave, and saw that:
>
> (a) all had RegSS/TotSS which cannot be negative,
>

Bob Ling reads badly.

Bob Ling is clueless when it comes to definitions
of R^2 that are useful for nonlinear regression -- or,
for non-OLS or -ML solutions to something that might
otherwise be called a linear regression problem, which
is the topic of this thread.... As Jerry wrote,

RegSS= 85, TSS= 82.5.

Thus, RegSS/TotSS is greater than 1.0, not "negative."

Thus, "1 minus the fraction" is what is negative.
[ ... ]

--
Rich Ulrich, wpilib@xxxxxxxx
http://www.pitt.edu/~wpilib/index.html
.