Re: R Square?
- From: "ajikoe@xxxxxxxxx" <ajikoe@xxxxxxxxx>
- Date: 11 Dec 2005 01:15:55 -0800
I have equations like:
x y ypred
1 11 12.879
2 5.5 123.5346
3 1 253.64
4 23 379.48
5 4 495.078
6 5 612.2066
where I get ypred from excel and add trendline I've got:
ypred = 0.7396x4 - 11.35x3 + 59.337x2 - 120.43x + 83.583
Rsquare = 0.3797
I want to know how Excel find it.
First I square the pearson correlation (this is define as function
rsq(x,y) in excel)
Instead of using rsq(x,y) I use rsq(y, ypred), I've got:
Rsquare = 0.00507954 (not the same as above)
Second I try to calculate Rsquare using
Rsq = 1 - SS(regression)/SS(total)
I've got:
Rsquare = -2594.624977 (not the same as above)
I wonder how excel trendline find the Rsquare ?
Best Regards,
pujo
.
- Follow-Ups:
- Re: R Square?
- From: Richard Ulrich
- Re: R Square?
- References:
- R Square?
- From: ajikoe@xxxxxxxxx
- Re: R Square?
- From: Richard Ulrich
- R Square?
- Prev by Date: EM algorithm for categorical data
- Next by Date: Re: R Square?
- Previous by thread: Re: R Square?
- Next by thread: Re: R Square?
- Index(es):
Relevant Pages
|
|