
LogLikelihoodFCN class for likelihood fits it is template to distinguish gradient and non-gradient case @ingroup FitMethodFunc
| virtual double | DoDerivative(const double* x, unsigned int icoord) const |
| virtual double | DoEval(const double* x) const |
| ROOT::Fit::LogLikelihoodFCN<ROOT::Math::IGradientFunctionMultiDim> | LogLikelihoodFCN<ROOT::Math::IGradientFunctionMultiDim>(const ROOT::Fit::LogLikelihoodFCN<ROOT::Math::IGradientFunctionMultiDim>&) |
| ROOT::Fit::LogLikelihoodFCN<ROOT::Math::IGradientFunctionMultiDim>& | operator=(const ROOT::Fit::LogLikelihoodFCN<ROOT::Math::IGradientFunctionMultiDim>&) |
| enum ROOT::Math::BasicFitMethodFunction | kUndefined | |
| kLeastSquare | ||
| kLogLikelihood | ||
| }; |
| const ROOT::Fit::UnBinData& | fData | |
| const ROOT::Fit::LogLikelihoodFCN<ROOT::Math::IGradientFunctionMultiDim>::IModelFunction& | fFunc | |
| vector<double> | fGrad | for derivatives |
| bool | fIsExtended | flag for indicating if likelihood is extended |
| unsigned int | fNEffPoints | number of effective points used in the fit |
| int | fWeight | flag to indicate if needs to evaluate using weight or weight squared (default weight = 0) |

clone the function (need to return Base for Windows)
{ return new LogLikelihoodFCN(fData,fFunc,fWeight,fIsExtended); }using BaseObjFunction::operator(); effective points used in the fit
{ return fNEffPoints; }i-th likelihood contribution and its gradient
Use sum of the weight squared in evaluating the likelihood (this is needed for calculating the errors)