class containg the result of the fit and all the related information (fitted parameter values, error, covariance matrix and minimizer result information) Contains a pointer also to the fitted (model) function, modified with the fit parameter values. When the fit is valid, it is constructed from a Minimizer and a model function pointer @ingroup FitMain
~FitResult() | |
double | Chi2() const |
double | Correlation(unsigned int i, unsigned int j) const |
double | CovMatrix(unsigned int i, unsigned int j) const |
int | CovMatrixStatus() const |
double | Edm() const |
double | Error(unsigned int i) const |
const vector<double>& | Errors() const |
ROOT::Fit::FitResult | FitResult() |
ROOT::Fit::FitResult | FitResult(const ROOT::Fit::FitConfig& fconfig) |
ROOT::Fit::FitResult | FitResult(const ROOT::Fit::FitResult&) |
ROOT::Fit::FitResult | FitResult(ROOT::Math::Minimizer& min, const ROOT::Fit::FitConfig& fconfig, const ROOT::Fit::FitResult::IModelFunction* f, bool isValid, unsigned int sizeOfData = 0, bool binFit = true, const ROOT::Math::IMultiGenFunction* chi2func = 0, unsigned int ncalls = 0) |
const ROOT::Fit::FitResult::IModelFunction* | FittedFunction() const |
void | GetConfidenceIntervals(const ROOT::Fit::BinData& data, double* ci, double cl = 0.95, bool norm = true) const |
void | GetConfidenceIntervals(unsigned int n, unsigned int stride1, unsigned int stride2, const double* x, double* ci, double cl = 0.95, bool norm = true) const |
void | GetCorrelationMatrix(TMatrixTSym<double>& mat) const |
void | GetCorrelationMatrix<TMatrixDSym>(TMatrixTSym<double>& mat) const |
void | GetCovarianceMatrix(TMatrixTSym<double>& mat) const |
void | GetCovarianceMatrix<TMatrixDSym>(TMatrixTSym<double>& mat) const |
const double* | GetErrors() const |
string | GetParameterName(unsigned int ipar) const |
const double* | GetParams() const |
double | GlobalCC(unsigned int i) const |
bool | HasMinosError(unsigned int i) const |
int | Index(const string& name) const |
bool | IsEmpty() const |
bool | IsParameterBound(unsigned int ipar) const |
bool | IsParameterFixed(unsigned int ipar) const |
bool | IsValid() const |
double | LowerError(unsigned int i) const |
double | MinFcnValue() const |
const string& | MinimizerType() const |
unsigned int | NCalls() const |
unsigned int | Ndf() const |
unsigned int | NFreeParameters() const |
bool | NormalizedErrors() const |
void | NormalizeErrors() |
unsigned int | NPar() const |
unsigned int | NTotalParameters() const |
ROOT::Fit::FitResult& | operator=(const ROOT::Fit::FitResult& rhs) |
double | Parameter(unsigned int i) const |
const vector<double>& | Parameters() const |
double | ParError(unsigned int i) const |
string | ParName(unsigned int i) const |
void | Print(ostream& os, bool covmat = false) const |
void | PrintCovMatrix(ostream& os) const |
double | Prob() const |
void | SetMinosError(unsigned int i, double elow, double eup) |
int | Status() const |
bool | Update(const ROOT::Math::Minimizer& min, bool isValid, unsigned int ncalls = 0) |
double | UpperError(unsigned int i) const |
double | Value(unsigned int i) const |
ROOT::Fit::FitResult::IModelFunction* | ModelFunction() |
void | SetModelFunction(ROOT::Fit::FitResult::IModelFunction* func) |
vector<unsigned int> | fBoundParams | list of limited parameters |
double | fChi2 | fit chi2 value (different than fval in case of chi2 fits) |
vector<double> | fCovMatrix | covariance matrix (size is npar*(npar+1)/2) where npar is total parameters |
int | fCovStatus | covariance matrix status code |
double | fEdm | expected distance from mimimum |
vector<double> | fErrors | errors |
ROOT::Fit::FitResult::IModelFunction* | fFitFunc | ! model function resulting from the fit. It is given by Fitter but it is managed by FitResult |
vector<unsigned int> | fFixedParams | list of fixed parameters |
vector<double> | fGlobalCC | global Correlation coefficient |
string | fMinimType | string indicating type of minimizer |
map<unsigned int,std::pair<double,double> > | fMinosErrors | map contains the two Minos errors |
unsigned int | fNCalls | number of function calls |
unsigned int | fNFree | number of fit free parameters (total parameters are in size of parameter vector) |
unsigned int | fNdf | number of degree of freedom |
bool | fNormalized | flag for indicating is errors are normalized |
vector<std::string> | fParNames | parameter names (only with FCN only fits, when fFitFunc=0) |
vector<double> | fParams | parameter values. Size is total number of parameters |
int | fStatus | minimizer status code |
double | fVal | minimum function value |
bool | fValid | flag for indicating valid fit |
Constructor from a fit-config for a dummy fit (e.g. when only one fcn evaluation is done)
Construct from a Minimizer instance after fitting Run also Minos if requested from the configuration
Update the fit result with a new minimization status To be run only if same fit is performed with same configuration Note that in this case MINOS is not re-run. If one wants to run also MINOS a new result must be created
minimization quantities minimizer type
{ return fMinimType; }
True if a fit result does not exist (even invalid) with parameter values
{ return (fParams.size() == 0); }
Return value of the objective function (chi2 or likelihood) used in the fit
{ return fVal; }
covariance matrix status code using Minuit convention : =0 not calculated, =1 approximated, =2 made pos def , =3 accurate
{ return fCovStatus; }
fitting quantities Return pointer to model (fit) function with fitted parameter values.
{ return fFitFunc; }
parameter values (return std::vector)
{ return fParams; }
parameter values (return const pointer)
{ return &fParams.front(); }
set the Minos errors for parameter i (called by the Fitter class when running Minos)
lower Minos error. If Minos has not run for parameter i return the parabolic error
upper Minos error. If Minos has not run for parameter i return the parabolic error
fill covariance matrix elements using a generic matrix class implementing operator(i,j) the matrix must be previously allocates with right size (npar * npar)
fill a correlation matrix elements using a generic symmetric matrix class implementing operator(i,j) the matrix must be previously allocates with right size (npar * npar)
get confidence intervals for an array of n points x. stride1 indicates the stride in the coordinate space while stride2 the stride in dimension space. For 1-dim points : stride1=1, stride2=1 for multi-dim points arranged as (x0,x1,...,xN,y0,....yN) stride1=1 stride2=n for multi-dim points arraged as (x0,y0,..,x1,y1,...,xN,yN,..) stride1=ndim, stride2=1 the confidence interval are returned in the array ci cl is the desired confidedence interval value norm is a flag to control if the intervals need to be normalized to the chi2/ndf value By default the intervals are corrected using the chi2/ndf value of the fit if a chi2 fit is performed
evaluate confidence interval for the point specified in the passed data sets the confidence interval are returned in the array ci cl is the desired confidence interval value
print the result and optionaly covariance matrix and correlations