LINEST, LOGEST, GROWTH or TREND??

[NlCO]

Hi all, I need to get the R2 statistic of a formula like this

f(t) = a + b*c^t

I've tried with LINEST(Range Y,LN(Range X),,TRUE) then the array three cells down and it should appear, but the numbers I get applies only to:

f(t) = b*c^t

Anybody knows how to get the real value or the R2 in any other way?

Thanks NlCO

[Jerry W. Lewis]

You have not said what is known and what must be estimated here. Based
on getting a result using your LINEST formula, I assume that c is known
and a,b, & t must be estimated. That would make the problem nonlinear
in the unknowns. Excel has no functionality to natively fit such
functions. You could use Solver to approximately minimize the sum of
squared deviations by changing trial values of a,b, & t.

Once you have values for a,b, & t, R2 would be
=DEVSQ(a+b*c_range^t)/DEVSQ(y_range)
which must be array entered (Ctrl+Shift+Enter) because of the
calculation in the numerator.

Jerry

NlCO wrote:

> Hi all, I need to get the R2 of a formula like this
>
> f(t) = a + b*c^t
>
> I've tried with LINEST(Range Y,LN(Range X),,TRUE) then the array three
> cells down and it should appear, but the numbers I get applies only
> to:
>
> f(t) = b*c^t
>
> Anybody knows how to get the real value or the r2 in any other way?
>
> Thanks NlCO