170 for (
UInt_t i=0; i<fEventSample.size(); i++)
delete fEventSample[i];
171 for (
UInt_t i=0; i<fForest.size(); i++)
delete fForest[i];
230 DeclareOptionRef(fGDTau=-1,
"GDTau",
"Gradient-directed (GD) path: default fit cut-off");
231 DeclareOptionRef(fGDTauPrec=0.01,
"GDTauPrec",
"GD path: precision of tau");
232 DeclareOptionRef(fGDPathStep=0.01,
"GDStep",
"GD path: step size");
233 DeclareOptionRef(fGDNPathSteps=10000,
"GDNSteps",
"GD path: number of steps");
234 DeclareOptionRef(fGDErrScale=1.1,
"GDErrScale",
"Stop scan when error > scale*errmin");
235 DeclareOptionRef(fLinQuantile,
"LinQuantile",
"Quantile of linear terms (removes outliers)");
236 DeclareOptionRef(fGDPathEveFrac=0.5,
"GDPathEveFrac",
"Fraction of events used for the path search");
237 DeclareOptionRef(fGDValidEveFrac=0.5,
"GDValidEveFrac",
"Fraction of events used for the validation");
239 DeclareOptionRef(fMinFracNEve=0.1,
"fEventsMin",
"Minimum fraction of events in a splittable node");
240 DeclareOptionRef(fMaxFracNEve=0.9,
"fEventsMax",
"Maximum fraction of events in a splittable node");
241 DeclareOptionRef(fNTrees=20,
"nTrees",
"Number of trees in forest.");
243 DeclareOptionRef(fForestTypeS=
"AdaBoost",
"ForestType",
"Method to use for forest generation (AdaBoost or RandomForest)");
244 AddPreDefVal(
TString(
"AdaBoost"));
245 AddPreDefVal(
TString(
"Random"));
247 DeclareOptionRef(fRuleMinDist=0.001,
"RuleMinDist",
"Minimum distance between rules");
248 DeclareOptionRef(fMinimp=0.01,
"MinImp",
"Minimum rule importance accepted");
250 DeclareOptionRef(fModelTypeS=
"ModRuleLinear",
"Model",
"Model to be used");
251 AddPreDefVal(
TString(
"ModRule"));
252 AddPreDefVal(
TString(
"ModRuleLinear"));
253 AddPreDefVal(
TString(
"ModLinear"));
254 DeclareOptionRef(fRuleFitModuleS=
"RFTMVA",
"RuleFitModule",
"Which RuleFit module to use");
255 AddPreDefVal(
TString(
"RFTMVA"));
256 AddPreDefVal(
TString(
"RFFriedman"));
258 DeclareOptionRef(fRFWorkDir=
"./rulefit",
"RFWorkDir",
"Friedman\'s RuleFit module (RFF): working dir");
259 DeclareOptionRef(fRFNrules=2000,
"RFNrules",
"RFF: Mximum number of rules");
260 DeclareOptionRef(fRFNendnodes=4,
"RFNendnodes",
"RFF: Average number of end nodes");
268 if (IgnoreEventsWithNegWeightsInTraining()) {
269 Log() << kFATAL <<
"Mechanism to ignore events with negative weights in training not yet available for method: "
270 << GetMethodTypeName()
271 <<
" --> please remove \"IgnoreNegWeightsInTraining\" option from booking string."
275 fRuleFitModuleS.ToLower();
276 if (fRuleFitModuleS ==
"rftmva") fUseRuleFitJF =
kFALSE;
277 else if (fRuleFitModuleS ==
"rffriedman") fUseRuleFitJF =
kTRUE;
278 else fUseRuleFitJF =
kTRUE;
282 else if (fSepTypeS ==
"giniindex") fSepType =
new GiniIndex();
283 else if (fSepTypeS ==
"crossentropy") fSepType =
new CrossEntropy();
286 fModelTypeS.ToLower();
287 if (fModelTypeS ==
"modlinear" ) fRuleFit.SetModelLinear();
288 else if (fModelTypeS ==
"modrule" ) fRuleFit.SetModelRules();
289 else fRuleFit.SetModelFull();
291 fPruneMethodS.ToLower();
296 fForestTypeS.ToLower();
297 if (fForestTypeS ==
"random" ) fUseBoost =
kFALSE;
298 else if (fForestTypeS ==
"adaboost" ) fUseBoost =
kTRUE;
299 else fUseBoost =
kTRUE;
304 if (fUseBoost && (!fUseRuleFitJF)) fTreeEveFrac = 1.0;
308 if (fTreeEveFrac<=0) {
311 fTreeEveFrac = min( 0.5, (100.0 +6.0*sqrt(
n))/
n);
314 VerifyRange(Log(),
"nTrees", fNTrees,0,100000,20);
315 VerifyRange(Log(),
"MinImp", fMinimp,0.0,1.0,0.0);
316 VerifyRange(Log(),
"GDTauPrec", fGDTauPrec,1
e-5,5
e-1);
317 VerifyRange(Log(),
"GDTauMin", fGDTauMin,0.0,1.0);
318 VerifyRange(Log(),
"GDTauMax", fGDTauMax,fGDTauMin,1.0);
319 VerifyRange(Log(),
"GDPathStep", fGDPathStep,0.0,100.0,0.01);
320 VerifyRange(Log(),
"GDErrScale", fGDErrScale,1.0,100.0,1.1);
321 VerifyRange(Log(),
"GDPathEveFrac", fGDPathEveFrac,0.01,0.9,0.5);
322 VerifyRange(Log(),
"GDValidEveFrac",fGDValidEveFrac,0.01,1.0-fGDPathEveFrac,1.0-fGDPathEveFrac);
323 VerifyRange(Log(),
"fEventsMin", fMinFracNEve,0.0,1.0);
324 VerifyRange(Log(),
"fEventsMax", fMaxFracNEve,fMinFracNEve,1.0);
326 fRuleFit.GetRuleEnsemblePtr()->SetLinQuantile(fLinQuantile);
327 fRuleFit.GetRuleFitParamsPtr()->SetGDTauRange(fGDTauMin,fGDTauMax);
328 fRuleFit.GetRuleFitParamsPtr()->SetGDTau(fGDTau);
329 fRuleFit.GetRuleFitParamsPtr()->SetGDTauPrec(fGDTauPrec);
330 fRuleFit.GetRuleFitParamsPtr()->SetGDTauScan(fGDTauScan);
331 fRuleFit.GetRuleFitParamsPtr()->SetGDPathStep(fGDPathStep);
332 fRuleFit.GetRuleFitParamsPtr()->SetGDNPathSteps(fGDNPathSteps);
333 fRuleFit.GetRuleFitParamsPtr()->SetGDErrScale(fGDErrScale);
334 fRuleFit.SetImportanceCut(fMinimp);
335 fRuleFit.SetRuleMinDist(fRuleMinDist);
341 Log() << kINFO <<
"" <<
Endl;
342 Log() << kINFO <<
"--------------------------------------" <<
Endl;
343 Log() << kINFO <<
"Friedmans RuleFit module is selected." <<
Endl;
344 Log() << kINFO <<
"Only the following options are used:" <<
Endl;
345 Log() << kINFO <<
Endl;
353 Log() << kINFO <<
"--------------------------------------" <<
Endl;
354 Log() << kINFO <<
Endl;
364 fRuleFit.UseImportanceVisHists();
366 fRuleFit.SetMsgType( Log().GetMinType() );
368 if (HasTrainingTree()) InitEventSample();
378 fMonitorNtuple=
new TTree(
"MonitorNtuple_RuleFit",
"RuleFit variables");
379 fMonitorNtuple->Branch(
"importance",&fNTImportance,
"importance/D");
380 fMonitorNtuple->Branch(
"support",&fNTSupport,
"support/D");
381 fMonitorNtuple->Branch(
"coefficient",&fNTCoefficient,
"coefficient/D");
382 fMonitorNtuple->Branch(
"ncuts",&fNTNcuts,
"ncuts/I");
383 fMonitorNtuple->Branch(
"nvars",&fNTNvars,
"nvars/I");
384 fMonitorNtuple->Branch(
"type",&fNTType,
"type/I");
385 fMonitorNtuple->Branch(
"ptag",&fNTPtag,
"ptag/D");
386 fMonitorNtuple->Branch(
"pss",&fNTPss,
"pss/D");
387 fMonitorNtuple->Branch(
"psb",&fNTPsb,
"psb/D");
388 fMonitorNtuple->Branch(
"pbs",&fNTPbs,
"pbs/D");
389 fMonitorNtuple->Branch(
"pbb",&fNTPbb,
"pbb/D");
390 fMonitorNtuple->Branch(
"soversb",&fNTSSB,
"soversb/D");
399 SetSignalReferenceCut( 0.0 );
403 fLinQuantile = 0.025;
406 fSepTypeS =
"GiniIndex";
407 fPruneMethodS =
"NONE";
408 fPruneStrength = 3.5;
423 if (Data()->GetNEvents()==0) Log() << kFATAL <<
"<Init> Data().TrainingTree() is zero pointer" <<
Endl;
428 fEventSample.push_back(
new Event(*
ev));
430 if (fTreeEveFrac<=0) {
432 fTreeEveFrac = min( 0.5, (100.0 +6.0*sqrt(
n))/
n);
434 if (fTreeEveFrac>1.0) fTreeEveFrac=1.0;
436 std::shuffle(fEventSample.begin(), fEventSample.end(), std::default_random_engine{});
438 Log() << kDEBUG <<
"Set sub-sample fraction to " << fTreeEveFrac <<
Endl;
448 if(!IsSilentFile()) InitMonitorNtuple();
451 this->InitEventSample();
459 fRuleFit.GetRuleEnsemblePtr()->ClearRuleMap();
469 if (IsNormalised()) Log() << kFATAL <<
"\"Normalise\" option cannot be used with RuleFit; "
470 <<
"please remove the option from the configuration string, or "
471 <<
"use \"!Normalise\""
487 fRuleFit.Initialize(
this );
493 Log() << kDEBUG <<
"Fitting rule coefficients ..." <<
Endl;
494 fRuleFit.FitCoefficients();
497 Log() << kDEBUG <<
"Computing rule and variable importance" <<
Endl;
498 fRuleFit.CalcImportance();
501 fRuleFit.GetRuleEnsemblePtr()->
Print();
505 Log() << kDEBUG <<
"Filling rule ntuple" <<
Endl;
506 UInt_t nrules = fRuleFit.GetRuleEnsemble().GetRulesConst().size();
509 rule = fRuleFit.GetRuleEnsemble().GetRulesConst(i);
510 fNTImportance =
rule->GetRelImportance();
511 fNTSupport =
rule->GetSupport();
512 fNTCoefficient =
rule->GetCoefficient();
513 fNTType = (
rule->IsSignalRule() ? 1:-1 );
514 fNTNvars =
rule->GetRuleCut()->GetNvars();
515 fNTNcuts =
rule->GetRuleCut()->GetNcuts();
516 fNTPtag = fRuleFit.GetRuleEnsemble().GetRulePTag(i);
517 fNTPss = fRuleFit.GetRuleEnsemble().GetRulePSS(i);
518 fNTPsb = fRuleFit.GetRuleEnsemble().GetRulePSB(i);
519 fNTPbs = fRuleFit.GetRuleEnsemble().GetRulePBS(i);
520 fNTPbb = fRuleFit.GetRuleEnsemble().GetRulePBB(i);
521 fNTSSB =
rule->GetSSB();
522 fMonitorNtuple->Fill();
525 fRuleFit.MakeVisHists();
526 fRuleFit.MakeDebugHists();
528 Log() << kDEBUG <<
"Training done" <<
Endl;
537 fRuleFit.InitPtrs(
this );
540 std::vector<const TMVA::Event*> tmp;
543 tmp.push_back(event);
545 fRuleFit.SetTrainingEvents( tmp );
549 rfAPI->WelcomeMessage();
554 Log() << kINFO <<
"Training ..." <<
Endl;
555 rfAPI->TrainRuleFit();
557 Log() << kDEBUG <<
"reading model summary from rf_go.exe output" <<
Endl;
558 rfAPI->ReadModelSum();
562 Log() << kDEBUG <<
"calculating rule and variable importance" <<
Endl;
563 fRuleFit.CalcImportance();
566 fRuleFit.GetRuleEnsemblePtr()->
Print();
568 if(!IsSilentFile())fRuleFit.MakeVisHists();
572 Log() << kDEBUG <<
"done training" <<
Endl;
581 fRanking =
new Ranking( GetName(),
"Importance" );
584 fRanking->AddRank(
Rank( GetInputLabel(
ivar), fRuleFit.GetRuleEnsemble().GetVarImportance(
ivar) ) );
595 fRuleFit.GetRuleEnsemble().AddXMLTo( parent );
603 fRuleFit.GetRuleEnsemblePtr()->ReadRaw(
istr );
611 fRuleFit.GetRuleEnsemblePtr()->ReadFromXML(
wghtnode );
622 return fRuleFit.EvalEvent( *GetEvent() );
631 Log() << kINFO <<
"Write monitoring ntuple to file: " << BaseDir()->GetPath() <<
Endl;
632 fMonitorNtuple->
Write();
641 fout <<
" // not implemented for class: \"" << className <<
"\"" << std::endl;
642 fout <<
"};" << std::endl;
643 fout <<
"void " << className <<
"::Initialize(){}" << std::endl;
644 fout <<
"void " << className <<
"::Clear(){}" << std::endl;
645 fout <<
"double " << className <<
"::GetMvaValue__( const std::vector<double>& inputValues ) const {" << std::endl;
646 fout <<
" double rval=" << std::setprecision(10) << fRuleFit.GetRuleEnsemble().GetOffset() <<
";" << std::endl;
647 MakeClassRuleCuts(
fout);
648 MakeClassLinear(
fout);
649 fout <<
" return rval;" << std::endl;
650 fout <<
"}" << std::endl;
651 fout << std::setprecision(
dp);
660 if (!fRuleFit.GetRuleEnsemble().DoRules()) {
661 fout <<
" //" << std::endl;
662 fout <<
" // ==> MODEL CONTAINS NO RULES <==" << std::endl;
663 fout <<
" //" << std::endl;
667 const std::vector< Rule* > *
rules = &(
rens->GetRulesConst());
670 std::list< std::pair<Double_t,Int_t> >
sortedRules;
676 fout <<
" //" << std::endl;
677 fout <<
" // here follows all rules ordered in importance (most important first)" << std::endl;
678 fout <<
" // at the end of each line, the relative importance of the rule is given" << std::endl;
679 fout <<
" //" << std::endl;
681 for ( std::list< std::pair<double,int> >::reverse_iterator
itpair =
sortedRules.rbegin();
687 fout <<
" if (" << std::flush;
695 if (
ic>0)
fout <<
"&&" << std::flush;
697 fout <<
"(" << std::setprecision(10) <<
valmin << std::flush;
698 fout <<
"<inputValues[" <<
sel <<
"])" << std::flush;
702 fout <<
"(inputValues[" <<
sel <<
"]" << std::flush;
703 fout <<
"<" << std::setprecision(10) <<
valmax <<
")" <<std::flush;
706 fout <<
") rval+=" << std::setprecision(10) << (*rules)[
ir]->GetCoefficient() <<
";" << std::flush;
709 fout << std::setprecision(
dp);
717 if (!fRuleFit.GetRuleEnsemble().DoLinear()) {
718 fout <<
" //" << std::endl;
719 fout <<
" // ==> MODEL CONTAINS NO LINEAR TERMS <==" << std::endl;
720 fout <<
" //" << std::endl;
723 fout <<
" //" << std::endl;
724 fout <<
" // here follows all linear terms" << std::endl;
725 fout <<
" // at the end of each line, the relative importance of the term is given" << std::endl;
726 fout <<
" //" << std::endl;
730 if (
rens->IsLinTermOK(
il)) {
736 << std::setprecision(10) <<
rens->GetLinCoefficients(
il)*
norm
737 <<
"*std::min( double(" << std::setprecision(10) <<
rens->GetLinDP(
il)
738 <<
"), std::max( double(inputValues[" <<
il <<
"]), double(" << std::setprecision(10) <<
rens->GetLinDM(
il) <<
")));"
758 Log() << col <<
"--- Short description:" <<
colres <<
Endl;
760 Log() <<
"This method uses a collection of so called rules to create a" <<
Endl;
761 Log() <<
"discriminating scoring function. Each rule consists of a series" <<
Endl;
762 Log() <<
"of cuts in parameter space. The ensemble of rules are created" <<
Endl;
763 Log() <<
"from a forest of decision trees, trained using the training data." <<
Endl;
764 Log() <<
"Each node (apart from the root) corresponds to one rule." <<
Endl;
765 Log() <<
"The scoring function is then obtained by linearly combining" <<
Endl;
766 Log() <<
"the rules. A fitting procedure is applied to find the optimum" <<
Endl;
767 Log() <<
"set of coefficients. The goal is to find a model with few rules" <<
Endl;
768 Log() <<
"but with a strong discriminating power." <<
Endl;
770 Log() << col <<
"--- Performance optimisation:" <<
colres <<
Endl;
772 Log() <<
"There are two important considerations to make when optimising:" <<
Endl;
774 Log() <<
" 1. Topology of the decision tree forest" <<
brk <<
Endl;
775 Log() <<
" 2. Fitting of the coefficients" <<
Endl;
777 Log() <<
"The maximum complexity of the rules is defined by the size of" <<
Endl;
778 Log() <<
"the trees. Large trees will yield many complex rules and capture" <<
Endl;
779 Log() <<
"higher order correlations. On the other hand, small trees will" <<
Endl;
780 Log() <<
"lead to a smaller ensemble with simple rules, only capable of" <<
Endl;
781 Log() <<
"modeling simple structures." <<
Endl;
782 Log() <<
"Several parameters exists for controlling the complexity of the" <<
Endl;
783 Log() <<
"rule ensemble." <<
Endl;
785 Log() <<
"The fitting procedure searches for a minimum using a gradient" <<
Endl;
786 Log() <<
"directed path. Apart from step size and number of steps, the" <<
Endl;
787 Log() <<
"evolution of the path is defined by a cut-off parameter, tau." <<
Endl;
788 Log() <<
"This parameter is unknown and depends on the training data." <<
Endl;
789 Log() <<
"A large value will tend to give large weights to a few rules." <<
Endl;
790 Log() <<
"Similarly, a small value will lead to a large set of rules" <<
Endl;
791 Log() <<
"with similar weights." <<
Endl;
793 Log() <<
"A final point is the model used; rules and/or linear terms." <<
Endl;
794 Log() <<
"For a given training sample, the result may improve by adding" <<
Endl;
795 Log() <<
"linear terms. If best performance is obtained using only linear" <<
Endl;
796 Log() <<
"terms, it is very likely that the Fisher discriminant would be" <<
Endl;
797 Log() <<
"a better choice. Ideally the fitting procedure should be able to" <<
Endl;
798 Log() <<
"make this choice by giving appropriate weights for either terms." <<
Endl;
800 Log() << col <<
"--- Performance tuning via configuration options:" <<
colres <<
Endl;
802 Log() <<
"I. TUNING OF RULE ENSEMBLE:" <<
Endl;
804 Log() <<
" " << col <<
"ForestType " <<
colres
805 <<
": Recommended is to use the default \"AdaBoost\"." <<
brk <<
Endl;
806 Log() <<
" " << col <<
"nTrees " <<
colres
807 <<
": More trees leads to more rules but also slow" <<
Endl;
808 Log() <<
" performance. With too few trees the risk is" <<
Endl;
809 Log() <<
" that the rule ensemble becomes too simple." <<
brk <<
Endl;
810 Log() <<
" " << col <<
"fEventsMin " <<
colres <<
brk <<
Endl;
811 Log() <<
" " << col <<
"fEventsMax " <<
colres
812 <<
": With a lower min, more large trees will be generated" <<
Endl;
813 Log() <<
" leading to more complex rules." <<
Endl;
814 Log() <<
" With a higher max, more small trees will be" <<
Endl;
815 Log() <<
" generated leading to more simple rules." <<
Endl;
816 Log() <<
" By changing this range, the average complexity" <<
Endl;
817 Log() <<
" of the rule ensemble can be controlled." <<
brk <<
Endl;
818 Log() <<
" " << col <<
"RuleMinDist " <<
colres
819 <<
": By increasing the minimum distance between" <<
Endl;
820 Log() <<
" rules, fewer and more diverse rules will remain." <<
Endl;
821 Log() <<
" Initially it is a good idea to keep this small" <<
Endl;
822 Log() <<
" or zero and let the fitting do the selection of" <<
Endl;
823 Log() <<
" rules. In order to reduce the ensemble size," <<
Endl;
824 Log() <<
" the value can then be increased." <<
Endl;
827 Log() <<
"II. TUNING OF THE FITTING:" <<
Endl;
829 Log() <<
" " << col <<
"GDPathEveFrac " <<
colres
830 <<
": fraction of events in path evaluation" <<
Endl;
831 Log() <<
" Increasing this fraction will improve the path" <<
Endl;
832 Log() <<
" finding. However, a too high value will give few" <<
Endl;
833 Log() <<
" unique events available for error estimation." <<
Endl;
834 Log() <<
" It is recommended to use the default = 0.5." <<
brk <<
Endl;
835 Log() <<
" " << col <<
"GDTau " <<
colres
836 <<
": cutoff parameter tau" <<
Endl;
837 Log() <<
" By default this value is set to -1.0." <<
Endl;
839 Log() <<
" This means that the cut off parameter is" <<
Endl;
840 Log() <<
" automatically estimated. In most cases" <<
Endl;
841 Log() <<
" this should be fine. However, you may want" <<
Endl;
842 Log() <<
" to fix this value if you already know it" <<
Endl;
843 Log() <<
" and want to reduce on training time." <<
brk <<
Endl;
844 Log() <<
" " << col <<
"GDTauPrec " <<
colres
845 <<
": precision of estimated tau" <<
Endl;
846 Log() <<
" Increase this precision to find a more" <<
Endl;
847 Log() <<
" optimum cut-off parameter." <<
brk <<
Endl;
848 Log() <<
" " << col <<
"GDNStep " <<
colres
849 <<
": number of steps in path search" <<
Endl;
850 Log() <<
" If the number of steps is too small, then" <<
Endl;
851 Log() <<
" the program will give a warning message." <<
Endl;
853 Log() <<
"III. WARNING MESSAGES" <<
Endl;
855 Log() << col <<
"Risk(i+1)>=Risk(i) in path" <<
colres <<
brk <<
Endl;
856 Log() << col <<
"Chaotic behaviour of risk evolution." <<
colres <<
Endl;
858 Log() <<
" The error rate was still decreasing at the end" <<
Endl;
859 Log() <<
" By construction the Risk should always decrease." <<
Endl;
860 Log() <<
" However, if the training sample is too small or" <<
Endl;
861 Log() <<
" the model is overtrained, such warnings can" <<
Endl;
862 Log() <<
" occur." <<
Endl;
863 Log() <<
" The warnings can safely be ignored if only a" <<
Endl;
864 Log() <<
" few (<3) occur. If more warnings are generated," <<
Endl;
865 Log() <<
" the fitting fails." <<
Endl;
866 Log() <<
" A remedy may be to increase the value" <<
brk <<
Endl;
868 << col <<
"GDValidEveFrac" <<
colres
869 <<
" to 1.0 (or a larger value)." <<
brk <<
Endl;
870 Log() <<
" In addition, if "
871 << col <<
"GDPathEveFrac" <<
colres
872 <<
" is too high" <<
Endl;
873 Log() <<
" the same warnings may occur since the events" <<
Endl;
874 Log() <<
" used for error estimation are also used for" <<
Endl;
875 Log() <<
" path estimation." <<
Endl;
876 Log() <<
" Another possibility is to modify the model - " <<
Endl;
877 Log() <<
" See above on tuning the rule ensemble." <<
Endl;
879 Log() << col <<
"The error rate was still decreasing at the end of the path"
881 Log() <<
" Too few steps in path! Increase "
882 << col <<
"GDNSteps" <<
colres <<
"." <<
Endl;
884 Log() << col <<
"Reached minimum early in the search" <<
colres <<
Endl;
886 Log() <<
" Minimum was found early in the fitting. This" <<
Endl;
887 Log() <<
" may indicate that the used step size "
889 Log() <<
" was too large. Reduce it and rerun." <<
Endl;
890 Log() <<
" If the results still are not OK, modify the" <<
Endl;
891 Log() <<
" model either by modifying the rule ensemble" <<
Endl;
892 Log() <<
" or add/remove linear terms" <<
Endl;
#define REGISTER_METHOD(CLASS)
for example
ROOT::Detail::TRangeCast< T, true > TRangeDynCast
TRangeDynCast is an adapter class that allows the typed iteration through a TCollection.
Option_t Option_t TPoint TPoint const char GetTextMagnitude GetFillStyle GetLineColor GetLineWidth GetMarkerStyle GetTextAlign GetTextColor GetTextSize void char Point_t Rectangle_t WindowAttributes_t Float_t Float_t Float_t Int_t Int_t UInt_t UInt_t Rectangle_t Int_t Int_t Window_t TString Int_t GCValues_t GetPrimarySelectionOwner GetDisplay GetScreen GetColormap GetNativeEvent const char const char dpyName wid window const char font_name cursor keysym reg const char only_if_exist regb h Point_t winding char text const char depth char const char Int_t count const char ColorStruct_t color const char Pixmap_t Pixmap_t PictureAttributes_t attr const char char ret_data h unsigned char height h Atom_t Int_t ULong_t ULong_t unsigned char prop_list Atom_t sel
Option_t Option_t TPoint TPoint const char GetTextMagnitude GetFillStyle GetLineColor GetLineWidth GetMarkerStyle GetTextAlign GetTextColor GetTextSize void char Point_t Rectangle_t WindowAttributes_t Float_t Float_t Float_t Int_t Int_t UInt_t UInt_t Rectangle_t Int_t Int_t Window_t TString Int_t GCValues_t GetPrimarySelectionOwner GetDisplay GetScreen GetColormap GetNativeEvent const char const char dpyName wid window const char font_name cursor keysym reg const char only_if_exist regb h Point_t winding char text const char depth char const char Int_t count const char ColorStruct_t color const char Pixmap_t Pixmap_t PictureAttributes_t attr const char char ret_data h unsigned char height h Atom_t Int_t ULong_t ULong_t unsigned char prop_list Atom_t Atom_t Atom_t Time_t type
Bool_t WriteOptionsReference() const
Implementation of the CrossEntropy as separation criterion.
Class that contains all the data information.
static void SetIsTraining(bool on)
Implementation of a Decision Tree.
Implementation of the GiniIndex as separation criterion.
Virtual base Class for all MVA method.
J Friedman's RuleFit method.
void GetHelpMessage() const override
get help message text
void ReadWeightsFromXML(void *wghtnode) override
read rules from XML node
void DeclareOptions() override
define the options (their key words) that can be set in the option string know options.
Double_t GetMvaValue(Double_t *err=nullptr, Double_t *errUpper=nullptr) override
returns MVA value for given event
void MakeClassLinear(std::ostream &) const
print out the linear terms
void TrainJFRuleFit()
training of rules using Jerome Friedmans implementation
void InitEventSample(void)
write all Events from the Tree into a vector of Events, that are more easily manipulated.
void MakeClassRuleCuts(std::ostream &) const
print out the rule cuts
void Init(void) override
default initialization
void InitMonitorNtuple()
initialize the monitoring ntuple
virtual ~MethodRuleFit(void)
destructor
void ReadWeightsFromStream(std::istream &istr) override
read rules from an std::istream
TTree * fMonitorNtuple
pointer to monitor rule ntuple
void ProcessOptions() override
process the options specified by the user
void AddWeightsXMLTo(void *parent) const override
add the rules to XML node
MethodRuleFit(const TString &jobName, const TString &methodTitle, DataSetInfo &theData, const TString &theOption="")
standard constructor
Bool_t HasAnalysisType(Types::EAnalysisType type, UInt_t numberClasses, UInt_t) override
RuleFit can handle classification with 2 classes.
const Ranking * CreateRanking() override
computes ranking of input variables
void MakeClassSpecific(std::ostream &, const TString &) const override
write specific classifier response
void Train(void) override
void WriteMonitoringHistosToFile(void) const override
write special monitoring histograms to file (here ntuple)
void TrainTMVARuleFit()
training of rules using TMVA implementation
Implementation of the MisClassificationError as separation criterion.
Ranking for variables in method (implementation)
A class describing a 'rule cut'.
J Friedman's RuleFit method.
Implementation of a rule.
Implementation of the SdivSqrtSplusB as separation criterion.
Timing information for training and evaluation of MVA methods.
Singleton class for Global types used by TMVA.
virtual Int_t Write(const char *name=nullptr, Int_t option=0, Int_t bufsize=0)
Write this object to the current directory.
virtual void Print(Option_t *option="") const
This method must be overridden when a class wants to print itself.
static TString Format(const char *fmt,...)
Static method which formats a string using a printf style format descriptor and return a TString.
A TTree represents a columnar dataset.
create variable transformations
MsgLogger & Endl(MsgLogger &ml)