I'm wondering if it is possible to have a fixed constant (y-interception) in LINEST ()?
If it does, where should I cast that magic spell in the function?
Solved: See Post #2 by MrShorty
I'm wondering if it is possible to have a fixed constant (y-interception) in LINEST ()?
If it does, where should I cast that magic spell in the function?
Solved: See Post #2 by MrShorty
Last edited by Wes-at-UoB; 06-01-2015 at 05:42 AM.
It is not a built in option in the LINEST() function. You need to reformulate your regression equation, then feed that reformulated equation into LINEST(). It looks something like:
1) normal regression equation for a straight line: y=m*x+b
2) subtract b from both sides: y-b=m*x
3) If needed to see how this works, substitute u=y-b: u=m*x
It should be obvious that u is a straight line function through the origin. By inputting your desired intercept as b into the above equation, you can then create a regression that can go through whatever y intercept you specify.
In Excel, since I am not like everyone else on these forums, I like to set this up with a helper column:Then feed column C as the known_y's parameter and column A as the known_x's parameter of the LINEST() function, being sure to make the constant function FALSE.Please Login or Register to view this content.
Originally Posted by shg
Oh MrShorty, first of all, thanks for introducing me the LINEST () function. It is very handy!
I am doing polynomial regression, ie Y = aX^2 + bX + c, which I think it should work the same as your suggestion. Right?
That is correct. And there may be other variations, too, for other conditions.
Thank you so much!
There are currently 1 users browsing this thread. (0 members and 1 guests)
Bookmarks