How can I made a non linear regression in ROOT?

ROOT
1

Hello everyone

I’m trying to fitting data making a non linear regression.
I have three variables and a constant.
Q,K,Y and So.
The data I know are Q,Y and So.
The result of the regression should be this.
Q=K * (Y^a) * (So^b)
How can I fit the K, a and b parameters using a non linear regression with known data? Coding in ROOT. Does it exist an specific library or something?

Thanks a lot

2

I guess @moneta can help you

3

Hi,

If you have a data set consisting of (So,Y, Q) data, where the predictors are Y and So and the depended variable Q , the simplest way in ROOT is to create a TGra2D class with (x,y) = (Y,So) and z = Q and you fit with a TF2 created with 3 parameters.
TF2 * f2 = new TF2(“f2”,"[K] * (x^[a]) * (y^[b])")

and then you can call TGraph2D::Fit.

An example is https://root.cern.ch/doc/master/graph2dfit_8C.html

Best Regards

Lorenzo

4

Thank you so much

It was very helpful to me. It works excellent.

Regards.

5

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.