Using the skew normal distribution or Owen's T function

Newbie
1

Hello,

I was looking into using either the skew normal distribution or Owen’s T function with TF1 and I found your page https://sft.its.cern.ch/jira/browse/ROOT-5106. From the page it looks like it should now be supported in ROOT’s TMath? If so, I am using ROOT v6.18/00 but I am having difficulty finding documentation on using these functions. Installing the Boost library and imitating steps found here, I was able to

root [0] #include <boost/math/special_functions/erf.hpp>
root [1] using namespace boost::math;
root [2] TF1 skewNormal("skewNormal", "pdf(skew_normal([0], [1], [2]), x)")
(TF1 &) Name: skewNormal Title: pdf(skew_normal([0], [1], [2]), x)

But is this the extent that the Boost library is used in ROOT?

2

Hi,

Your code will work fine if you include the right header file from Boost and you have it in your include path.
For example:

#include <boost/math/distributions/skew_normal.hpp>
using namespace boost::math;
TF1 skewNormal("skewNormal", "pdf(skew_normal([0], [1], [2]), x)",-3,3);
skewNormal.SetParameters(0,1,2); 
skewNormal.Draw();

Or in the case of the Owens T function:

#include "boost/math/special_functions/owens_t.hpp"
using namespace boost::math;
TF2 owensT("owenst","owens_t(x,y)",-5,5,-5,5);
owensT.Draw("CONTZ");

As far you include the correct files that can be parsed by Cling you can use any external functions.
In some cases if the implementation is in a library, you might need to load that library

Lorenzo