Re: [TMVA-users] Training a BDT with sweights

[Christophe C. <chr...@tu...>]

Hi,

we tried exchanging the Tools files you supplied in the ROOT v5.34.03 sources. We exchanged

tmva/src/Tools.cxx and
tmva/inc/TMVA/Tools.h

via the files you supplied. Afterwards, we configured ROOT via

./configure --enable-mathmore --enable-memstat --enable-minuit2 --enable-python --enable-roofit --enable-tmva --enable-gsl-shared --enable-afs

The make afterwards fails with:

bin/rmkdepend -R -ftmva/src/G__TMVA2.d -Y -w 1000 -- -pipe -m64 -Wshadow -Wall -W -Woverloaded-virtual -fPIC -Iinclude -pthread -D__cplusplus -Icint/cint/lib/prec_stl -Icint/cint/stl -I/usr/local/root-v5-34-03/cint/cint/inc -- tmva/src/G__TMVA2.cxx
g++ -pipe -m64 -Wshadow -Wall -W -Woverloaded-virtual -fPIC -Iinclude -pthread -I. -I/usr/local/root-v5-34-03/cint/cint/inc -o tmva/src/G__TMVA2.o -c tmva/src/G__TMVA2.cxx
tmva/src/G__TMVA2.cxx: In function ‘int G__G__TMVA2_392_0_12(G__value*, const char*, G__param*, int)’:
tmva/src/G__TMVA2.cxx:7737: error: ISO C++ forbids declaration of ‘type name’ with no type
tmva/src/G__TMVA2.cxx:7737: error: template argument 1 is invalid
tmva/src/G__TMVA2.cxx:7737: error: template argument 2 is invalid
tmva/src/G__TMVA2.cxx:7737: error: expected primary-expression before ‘long’
tmva/src/G__TMVA2.cxx:7737: error: expected `)' before ‘long’
tmva/src/G__TMVA2.cxx:7741: error: ISO C++ forbids declaration of ‘type name’ with no type
tmva/src/G__TMVA2.cxx:7741: error: template argument 1 is invalid
tmva/src/G__TMVA2.cxx:7741: error: template argument 2 is invalid
tmva/src/G__TMVA2.cxx:7741: error: expected primary-expression before ‘long’
tmva/src/G__TMVA2.cxx:7741: error: expected `)' before ‘long’
make: *** [tmva/src/G__TMVA2.o] Error 1

To me it seems something in ROOT's automated dictionary building goes wrong, but I don't understand much of it. We are using gcc 4.3.5 coming with the LHCb software to build ROOT.

Cheers,

Florian & Christophe



Am Freitag, 9. November 2012 um 17:31 schrieb Helge Voss:

> Hi Chrsitophe,
>  
> it should actually be enough to copy use the attached
> Tools.h and Tools.cxx and replace the "old" ones....
> (please let me know if you shoudl have any problems using
> those)
>  
>  
> otherwise, the full current development version you get via:
>  
>  
> svn co https://root.cern.ch/svn/root/branches/dev/tmva
>  
>  
>  
> 2012/11/9 Christophe Cauet <chr...@tu... (mailto:chr...@tu...)>:
> > Hi Helge,
> >  
> > thanks for the quick reply! I only found a quite old tmva version in the
> > repo http://tmva.svn.sourceforge.net/viewvc/tmva/
> >  
> > Can you point me to a newer one? Or, if its not too much work for you, you
> > may just send me the patch?
> >  
> > Thanks
> > Christophe
> >  
> >  
> > Am Freitag, 9. November 2012 um 16:21 schrieb Helge Voss:
> >  
> > Hi Christophe,
> >  
> > yes.. that's a tiny bug I've fixed already some time ago, but it
> > didn't make it's
> > way yet into the official ROOT release. Its caused by the fact that the
> > weights
> > are not taken into account when calculating if the mean value of
> > MVA_Signal > or < than mean MVA_Backgound... hence deciding if a cut
> > selects rather Signal or Background.
> >  
> > It wasn't spotted earlier, as typcially the event weight's don't
> > change the distributions
> > so much... but with s-weights, they do. Hence... Training and
> > everything is correct,
> > it is just when doing the efficinecy plots/calculations that this
> > matters and goes wrong.
> >  
> >  
> >  
> > Now, unfortunatly, the current 'dev' version that you could check out
> > from svn is quite a
> > bit different fromt ROOT version. I you don't mind that.. please use
> > that. Otherwise I
> > would have to get you a patch.. and you'd have to 're-compile' your ROOT..
> >  
> > sorry..
> >  
> > Cheers,
> >  
> > Helge
> >  
> >  
> > 2012/11/9 Christophe Cauet <chr...@tu... (mailto:chr...@tu...)>:
> >  
> > Dear experts,
> >  
> > I'm experiencing some trouble training a BDT with TMVA.
> >  
> > As far as I can decide the classifier distribution looks not too bad for the
> > BDT
> > * https://dl.dropbox.com/u/171315/tmva/classifier_distribution_BDT.png
> > as well as for a Fisher discriminant which I use as a reference, since I was
> > always getting quite good results with it:
> > * https://dl.dropbox.com/u/171315/tmva/classifier_distribution_Fisher.png
> >  
> > However the the cut efficiencies as well as the ROC curve look very strange:
> > * https://dl.dropbox.com/u/171315/tmva/cut_efficiencies_BDT.png
> > * https://dl.dropbox.com/u/171315/tmva/cut_efficiencies_Fisher.png
> > * https://dl.dropbox.com/u/171315/tmva/ROC_curve.png
> >  
> > I'm using a single TTree as input and sweights from an sSplot via:
> >  
> > factory->SetWeightExpression("sweight_sig", "Signal");
> > factory->SetWeightExpression("sweight_bkg", "Background");
> >  
> > to discriminate between signal and background events for training and
> > testing.
> >  
> > There are about 130k signal and 20k background events in my sample.
> >  
> > As I never experienced such problems before I guess it has something to do
> > with the sweights.
> >  
> > Do you ever tried that before or have a clue whats going wrong?
> >  
> > Thanks
> > Christophe
> >  
> >  
> >  
> > PS:
> > The TMVA output is:
> > https://dl.dropbox.com/u/171315/tmva/tmva_output.txt
> >  
> > The factory options are
> > V:!Silent:Color:DrawProgressBar:Transformations=I;D;P;G,D:AnalysisType=Classification
> >  
> > The split options are:
> > SplitMode=Random:!V:SplitSeed=0:nTrain_Signal=0:nTest_Signal=0:nTrain_Background=0:nTest_Background=0
> >  
> > The BDT options are:
> > !V:NTrees=400:nEventsMin=400:MaxDepth=3:BoostType=AdaBoost:SeparationType=GiniIndex:nCuts=20:PruneMethod=NoPruning:VarTransform=Decorrelate
> > I also tried adding IgnoreNegWeightsInTraining
> >  
> > ------------------------------------------------------------------------------
> > Everyone hates slow websites. So do we.
> > Make your web apps faster with AppDynamics
> > Download AppDynamics Lite for free today:
> > http://p.sf.net/sfu/appdyn_d2d_nov
> > _______________________________________________
> > TMVA-users mailing list
> > TMV...@li... (mailto:TMV...@li...)
> > https://lists.sourceforge.net/lists/listinfo/tmva-users
> >  
>  
>  
>  
>  
>  
> Anhänge:  
> - Tools.cxx
>  
> - Tools.h
>