Hello,
I’m trying to make a 2D adaptive kernel estimate for two variables I have in a tree. One is an energy (“e”), and it happens that any sensible parametrization will be in the log of the energy. So I want to use RooNDKeysPdf with the variables “s” and “log(e)”. But as far as I can tell, RooDataSet’s constructor does not accept formulas (and neither does its createHistogram method, obscuring the display), and when (following the tutorial) I try to use a formula (RooNDKeysPdf kest4("kest4","kest4",RooArgSet(var_s,lne),*data2,"amv"))
instead of a plain variable (RooNDKeysPdf kest4("kest4","kest4",RooArgSet(var_s,var_e),*data2,"am") ,
with:
RooRealVar var_s( "s", "s", 0, 3 );
RooRealVar var_e( "e", "e", 20, 320 );
RooDataSet* data2 = new RooDataSet( "data2", "data2", RooArgList( var_s, var_e, var_eta ), Import( tr ),Cut( cut ) ) ;
RooFormulaVar lne( "lne", "ln(e)", "log(e)", RooArgList( var_e ) );
)
I get an immediate (that is before the verbose printing) segmentation violation.
Is this doable?
Any ideas what I’m doing wrong?
thanks,
Amnon