I also get the same 6th degree equation with LINEST function.
I get it from 13 x values and want to use it to calculate 200 values in between these 13 values. I get way high results.
Thanks in advance,
I also get the same 6th degree equation with LINEST function.
I get it from 13 x values and want to use it to calculate 200 values in between these 13 values. I get way high results.
Thanks in advance,
Last edited by kameyji; 12-09-2023 at 05:55 PM.
I don't see any 6th order polynomial regressions in any of the LINEST() functions I could find, nor in any of the charts I could find. Can you help us find these erroneous 6th order polynomial calculations?
Originally Posted by shg
Sorry, i uploaded the old attachment, now i've edited it.
The regression equation being used in the LINEST() function is y=ax^6+bx^5+cx^4+dx^3+ex^2+fx+g. I notice that, in the formula in D21..., the equation entered is y=ax^6-bx^5+cx^4-dx^3+ex^2+fx+g -- effectively changing the sign on the x^5 and x^3 terms. Make sure that the formula you enter in D21... is the exact same as the regression equation being used by the LINEST() function. If it helps, I usually prefer to use the SERIESSUM() function for high order polynomials (https://support.microsoft.com/en-us/...rs=en-us&ad=us ) =$M$20+SERIESSUM(C21,6,-1,$G$20:$L$20).
I checked again sir but, both x^5 and x^3 terms coefficient is minus. Can i just SERIES function between the 13 values?
Thanks,
Yes, the coefficients are negative, but, assuming I understand the regression equation you are trying to use, you want to add the negative, not subtract the negative. There's a big difference between:
(4e-9)*x^6 + (-1.5E-6)*x^5 + ...
and
(4e-9)*x^6 - (-1.5e-6)*x^5 + ...
The former is what I would expect, and the latter is what you have.
Okay, i see now. Such simple mistake...
Thank you again sir.
There are currently 1 users browsing this thread. (0 members and 1 guests)
Bookmarks