Is there a chebyshev (or alternative) allocation routine?

Hello,

Is there a tool available in ROOT for finding the optimal allocation of sampling points given a function? I know this can be done with several methods, such as chebyshev nodes, but I could use any method that is available in ROOT. Is there such a method?

Hi,

What would you like exactly to do ? Approximate a function using polynomials ?
If this is what you are looking for, there is a class in ROOT for doing it. It is ROOT::Math::ChebyshevApprox,
project-mathlibs.web.cern.ch/pro … pprox.html
in the MathMore library and base don GSL

Best Regards

Lorenzo

[quote=“moneta”]Hi,

What would you like exactly to do ? Approximate a function using polynomials ?
If this is what you are looking for, there is a class in ROOT for doing it. It is ROOT::Math::ChebyshevApprox,
project-mathlibs.web.cern.ch/pro … pprox.html
in the MathMore library and base don GSL

Best Regards

Lorenzo[/quote]

No, I don’t need to approximate a function.

What I am looking for is a way to optimize sampling points. For example, if I have a function on some domain that I need to sample, but I only have a finite number of points to sample with, I want to find where to place those points to maximize the amount of information gathered. I have a well-behaved estimator function F with all continuous derivatives. I am looking for a way to allocate N sampling points in the domain based on the estimator function.

I am wondering if any part of this method is in the root libraries somewhere.

Hi,
We then don;t have such tool in ROOT. You could try maybe to see if something exists in R. You can now call R functions from ROOT using the ROOT-R interface, see

root.cern.ch/drupal/content/how- … -interface

Lorenzo