Fitting Power function and to record R-square

[drniranjan]

Hi,

I have the data sets:

A B C D E F so on...
1 1 3 4 4 5
2 4 1 2 4 1
2 8 2 7 2 9
1 6 2 9 3 9

I need the R squareof the fitting power function to the consequtive data pairs...means, between A &B; C &D; E and F; .....so on.........

finally to record the R square in the bottom of the 1st column of each pair...

regards, N <!-- / message -->

 

[drniranjan]

Hi,

I have the data sets:

A B C D E F so on...
1 1 3 4 4 5
2 4 1 2 4 1
2 8 2 7 2 9
1 6 2 9 3 9

I need the R squareof the fitting power function to the consequtive data pairs...means, between A &B; C &D; E and F; .....so on.........

finally to record the R square in the bottom of the 1st column of each pair...

regards, N <!-- / message -->
<!-- / message -->

 

[brettdj]

To fit a power series with a R2
y = Ax^b

You can use natural logs to solve with linest

If in A6 to B8 you enter this array formula
=LINEST(LN(B2:B5),LN(A2:A5),,TRUE)


you will get
1.5000 0.8959
1.31 0.64
0.396326066 0.907353703

A6 is your power, b, (1.50)
EXP(B6) = EXP(.8959) = 2.44 is a

so y= 2.44x^1.50
and R2 is in A8, ie 0.39

Cheers

Dave

Repeat for your other column pairs

 

[drniranjan]

hi,
Thanks for the help...formulae is working fine

 

[mirabeau]

When you ask for R-square, do you mean the squared correlation between y and Ax^b, or do you mean the squared correlation between Ln and ln(A)+b*Ln(x)?

These are not the same.

 

[drniranjan]

hi,
I meant the goodness of fitting coefficent.......while fitting to any equation we use to calculate gof indices to see the performace of the model (equation)...........so R2 is one of the few criteria to see the fitting performace............means how best the equation produces Y, which is very close to the actual Y.........

regards, niranjan